New Linode competitor - Digitalocean
Digitalocean.com
I'm not looking to put Linode down, they have been incredibly reliable and helpful, I'm just trying to make an informed decision.
223 Replies
A: No we do not limit monthly bandwidth in anyway. We allow you to fully utilize the network and burst traffic without limits.
Yeah, that always works out well because nobody would ever think of abusing that and hogging all the bandwidth for themselves.
@deadwalrus:
It does look fairly legit from a glance … would love to see more reviews. The pricing is quite attractive if it proves reliable …
Yes, they're definitely legitimate - See below.
in fact, I would be more worried than in 6 months or so they would get sold out to some other company.
i would also be weary as they're hitting all the keywords (cloud, ssd, unlimited bw, pay-as-you-grow, etc)
@vonskippy:
Q: Do you limit bandwidth? A: No we do not limit monthly bandwidth in anyway. We allow you to fully utilize the network and burst traffic without limits.
Yeah, that always works out well because nobody would ever think of abusing that and hogging all the bandwidth for themselves.
Yes, I agree. One of the owners does address this in the comments on Hackernews - http://news.ycombinator.com/item?id=5059723
@kbar:
imo just because they're sponsored by a vc doesn't mean they're legit.
in fact, I would be more worried than in 6 months or so they would get sold out to some other company
Depends on what you mean by 'legit'. I only provided the links to add some credibility and background. imho they are less likely to be 'fly-by-nighters' or deceptive/dishonest with a big name behind them but its certainly possible and I would be cautious myself.
$5 per month for a VPS with SSDs and unmetered gigabit connections?
No, that's not legit. Even OVH isn't selling unmetered gigabit connections anywhere remotely that cheaply, and OVH's entire purpose is to be disruptive to the market by leveraging their massive size to undercut the competition.
Looking at some actual benchmarks, their performance is pretty poor too. 8MB/s on SSDs? Yikes.
Selling unmetered gigabit connections for $5 per month means you won't be in business for long.
EDIT: This sort of offering is practically the definition of fly-by-night, because you're selling your product so far below cost.
I get ~6.5MB/s downloading the Newark Linode speed test file to a 'NYC' droplet.
And the disk looks good to me -
dd if=/dev/zero of=/tmp/foo bs=1024k count=10240
10240+0 records in
10240+0 records out
10737418240 bytes (11 GB) copied, 42.8255 s, 251 MB/s
(/tmp is not mounted on RAM)
Delivered: 52 megabits per second.
Beyond the obvious problem with the throughput, benchmarks aren't necessarily the concern, so much as the whole "impossible to make a profit if they deliver what they promise" issue.
@Guspaz:
Promised: 1000 megabits per second.
Delivered: 52 megabits per second.
Beyond the obvious problem with the throughput, benchmarks aren't necessarily the concern, so much as the whole "impossible to make a profit if they deliver what they promise" issue.
It's a marketing tactic: Offer steep discounts to build up a reputation then start increasing prices as they become a known commodity. Pretty typical, particularly for Y-combinator ventures.
@deadwalrus:
It's a marketing tactic: Offer steep discounts to build up a reputation then start increasing prices as they become a known commodity. Pretty typical, particularly for Y-combinator ventures.
If so, it's a BAD marketing tactic for their business model.
Any knowledgeable VPS host would know that saying "Unlimited" is the sure mark of either scam or incompetence.
> If so, it's a BAD marketing tactic for their business model.
Any knowledgeable VPS host would know that saying "Unlimited" is the sure mark of either scam or incompetence.
I dunno about that. I mean, we're talking about it.
I would consider switching if they prove to provide a consistent service. To be honest, we haven't seen much in the way of lowered prices or hardware upgrades around here in the last few years. Last RAM upgrade was summer 2010, and I wouldn't consider a 40% disk space increase over the last two years anything to jump up and shout about. It's starting to become noticeable on the RAM / disk front.
It's not clear what exactly they're promising. They say "All servers come with 1Gb/sec. network interface." Does that mean every cloud server, or every physical host? If it's every cloud server, how do they fit that many gigabit interfaces into a single physical host? If it's every physical host, then how do they make any kind of guarantee about performance?
They're selling these things as SSD-based. I've never seen SSD pricing, individually or in bulk, go much below $0.50 per gigabyte (and enterprise-grade is often much higher), so at $5/mth, it would take them two months just to cover the cost of storage, let alone the cost of bandwidth or the servers themselves…
Their benchmarks from what I've seen are already very low-end, and bandwidth performance is already rather lackluster. I'd much rather be on a platform like Linode that promises me a reliable amount of performance in a fixed quantity (like my 200GB of bandwidth) rather than a platform like DigitalOcean that says "free bandwidth" but doesn't really say if I'll get 1 megabit or 1000 megabits per second.
A) That's how I pay for my utilities. I don't like it now. I have never liked it. I like flat rate.
B) I like the people at Linode.
C) I like real disks. Not saying SSD is flaky, SSD's are still new, this too:
@Guspaz:
Linode might be a bit behind the curve in terms of passing on savings through dropping costs, but DigitalOcean is selling services at "scam warning bell" pricepoints. It's impossible to sell and deliver unmetered gigabit connections for $5 and not take huge losses. Either performance must be degraded (best-effort type scenario) or the price must be raised.
It's not clear what exactly they're promising. They say "All servers come with 1Gb/sec. network interface." Does that mean every cloud server, or every physical host? If it's every cloud server, how do they fit that many gigabit interfaces into a single physical host? If it's every physical host, then how do they make any kind of guarantee about performance?
They're selling these things as SSD-based. I've never seen SSD pricing, individually or in bulk, go much below $0.50 per gigabyte (and enterprise-grade is often much higher), so at $5/mth, it would take them two months just to cover the cost of storage, let alone the cost of bandwidth or the servers themselves…
Their benchmarks from what I've seen are already very low-end, and bandwidth performance is already rather lackluster. I'd much rather be on a platform like Linode that promises me a reliable amount of performance in a fixed quantity (like my 200GB of bandwidth) rather than a platform like DigitalOcean that says "free bandwidth" but doesn't really say if I'll get 1 megabit or 1000 megabits per second.
They really do need to tighten their sales copy but I'm pretty sure they mean a gigabit interface per physical server. You seem to know quite a bit about what things cost in this industry, would you be able to roughly estimate their cost for a single 512/20GB, assuming they are leasing their hardware in bulk at a good price? I have no idea what sort of markup is typical in this industry but given the competition I would guess it's very low which is what really sets off my warning bells. Anyway, I got $20 free voucher with the code ""SSDBEAR20" which was published on hackernews and I'm setting up a bunch of TOR exit nodes to see how they cope.
I had a server up there for a couple of weeks running as an online backup clone of one that I have at Linode, they suspended my account and deleted the server saying it had been compromised.
When I asked them for details and to restore it they did restore it (took a while) but couldn't explain about the server being compromised. In the end they said it was using too many resources and because my account was fairly new they thought I was a spammer.
They credited me $10 for my trouble, not something I actually care about I'd rather they'd emailed me warning me about the server before deleting it, shortly after this incident I deleted all my images from there and I'll use someone else for online backup clones. Luckily I didn't have anything critical there.
With any new hosting company try them out for a while before you think about using them. I've not found anyone that has support and services as good as Linode which is why I'm happy to pay what I pay for what I get.
@obs:
I had a server up there for a couple of weeks running as an online backup clone of one that I have at Linode, they suspended my account and deleted the server saying it had been compromised.
Thanks, that's all I need to hear to stick with Linode
A few interesting things came out:
* There is no private network. This is still being beta-tested. For anyone with multiple instances that communicate with each other, this is a no-no. I have no idea if the private network will be secure like that at Linode or just another open LAN.
* They are not willing to give any form of guarantee regarding speed, be it on network, processor or disk. They say their guarantee is that the better their service, the better their business. Well, that's all fine and dandy unless their target market is start-ups that don't really care about performance yet. Then it matters a lot. I told them my concern was a "droplet" that slowed over time as more resources were used with a growing customer base. Their response was a $10 credit to the account so that I could test it for longer. Thanks, but I'm thinking more of a year or two from now when presumably their numbers are a lot higher.
* They are obsessed about obtaining my credit-card. You have this trial-period of a number of hours and if not given, the account is suspended. To re-activate it you need to enter your credit-card or make a pay-pal payment. Very pushy. That prompts me to take a step back.
I have not tested their system because, bottom-line, I don't like the smell of it.
@nickdan:
I have not tested their system because, bottom-line, I don't like the smell of it.
"fishy"
heuahuuheuhuehuhueu
@Benio:
They really do need to tighten their sales copy but I'm pretty sure they mean a gigabit interface per physical server. You seem to know quite a bit about what things cost in this industry, would you be able to roughly estimate their cost for a single 512/20GB, assuming they are leasing their hardware in bulk at a good price? I have no idea what sort of markup is typical in this industry but given the competition I would guess it's very low which is what really sets off my warning bells. Anyway, I got $20 free voucher with the code ""SSDBEAR20" which was published on hackernews and I'm setting up a bunch of TOR exit nodes to see how they cope.
More like I've just bought my fair share of SSDs and know they're not cheap. For enterprise drives, I'd expect to pay at least $2/GB. I'm seeing Intel's DC S3700 (which is probably the best enterprise SSD on the market before you start going to stuff like FusionIO) for $2.65/GB.
I've seen consumer drives like the Intel 330 go down, when on sale with a mail in rebate, as low as $0.50 per gig.
No idea how bulk pricing would affect any of this.
@Benio:
What makes Linode so much better that they are worth 4x the amount? What made me come in the first place: a track record of reliability and quality, with promised resources delivered.
@nickdan:
There is no private network. This is still being beta-tested. For anyone with multiple instances that communicate with each other, this is a no-no. I have no idea if the private network will be secure like that at Linode or just another open LAN.
To be fair, the only purpose of a private network is to avoid getting charged for internal bandwidth; if there is no bandwidth charges, a private network is pointless. You shouldn't be trusting a private network to be secure anyhow, you should be using a VPN.
FYI, Linode's private network is just another open LAN; you'll get broadcast packets from other peoples' linodes.
@Guspaz:
@nickdan:There is no private network. This is still being beta-tested. For anyone with multiple instances that communicate with each other, this is a no-no. I have no idea if the private network will be secure like that at Linode or just another open LAN.
To be fair, the only purpose of a private network is to avoid getting charged for internal bandwidth; if there is no bandwidth charges, a private network is pointless. You shouldn't be trusting a private network to be secure anyhow, you should be using a VPN.
FYI, Linode's private network is just another open LAN; you'll get broadcast packets from other peoples' linodes.
That's not my understanding of the matter. See the post by Caker in this thread:. I also confirmed this with Linode Support. I'll be very disappointed to learn it is not the case. http://forum.linode.com/viewtopic.php?p=39528
@nickdan:
@Guspaz:To be fair, the only purpose of a private network is to avoid getting charged for internal bandwidth; if there is no bandwidth charges, a private network is pointless. You shouldn't be trusting a private network to be secure anyhow, you should be using a VPN.
FYI, Linode's private network is just another open LAN; you'll get broadcast packets from other peoples' linodes.
That's not my understanding of the matter. See the post by Caker in this thread:. I also confirmed this with Linode Support. I'll be very disappointed to learn it is not the case. http://forum.linode.com/viewtopic.php?p=39528
Guspaz was talking about broadcast traffic, while I'm pretty sure Caker's comments in that thread are in regards to unicast traffic.
Think of the private network as a big LAN switch. While your Linode's private interface will only receive unicast traffic directed at it, everyone sees broadcast traffic. There are ingress controls so no-one can spoof your private address (but I'd expect that on the public interface too), but anyone who generates unicast traffic at your Linode can reach it. I'm not sure what other sort of configuration you'd expect to make sense. Unless you're expecting Linode to define and maintain individual VLANs at layer 2 for all customers, but I'd expect that to be unreasonable (and unnecessary) in terms of scale and maintenance.
In terms of security, treat it just like the public interface, and filter to just those private addresses that you know belong to your Linodes. No differently than your public interface, or IPv6 traffic (which has no "private" network).
– David
@db3l:
@nickdan:
@Guspaz:To be fair, the only purpose of a private network is to avoid getting charged for internal bandwidth; if there is no bandwidth charges, a private network is pointless. You shouldn't be trusting a private network to be secure anyhow, you should be using a VPN.
FYI, Linode's private network is just another open LAN; you'll get broadcast packets from other peoples' linodes.
That's not my understanding of the matter. See the post by Caker in this thread:. I also confirmed this with Linode Support. I'll be very disappointed to learn it is not the case. http://forum.linode.com/viewtopic.php?p=39528
Guspaz was talking about broadcast traffic, while I'm pretty sure Caker's comments in that thread are in regards to unicast traffic.Think of the private network as a big LAN switch. While your Linode's private interface will only receive unicast traffic directed at it, everyone sees broadcast traffic. There are ingress controls so no-one can spoof your private address (but I'd expect that on the public interface too), but anyone who generates unicast traffic at your Linode can reach it. I'm not sure what other sort of configuration you'd expect to make sense. Unless you're expecting Linode to define and maintain individual VLANs at layer 2 for all customers, but I'd expect that to be unreasonable (and unnecessary) in terms of scale and maintenance.
In terms of security, treat it just like the public interface, and filter to just those private addresses that you know belong to your Linodes. No differently than your public interface, or IPv6 traffic (which has no "private" network).
– David
Thanks for taking the time to explain, David. I had the wrong idea of things.
Linode gets knocked around for lack of transparency (which I don't find to be the case) but there are few vendors that allow open discourse about their competitors on their own company forum.
@vonskippy:
It would be interesting to know if Digital Ocean would allow this thread on their discussion forum.
Linode gets knocked around for lack of transparency (which I don't find to be the case) but there are few vendors that allow open discourse about their competitors on their own company forum.
Yep, skippy excellent insight.
Linode - Are you going to take this? How about lowering your costs for a while to keep these upstarts from stealing your customers?
@sednet:
Linode - Are you going to take this? How about lowering your costs for a while to keep these upstarts from stealing your customers?
If the "upstarts" raise their prices people will come back if they left, or more likely when something breaks and it takes ages to resolve instead of Linode's normal quick fix time then they'll come back.
@sednet:
Linode - Are you going to take this? How about lowering your costs for a while to keep these upstarts from stealing your customers?
If I understand Linode's core business model correctly. it is to offer high quality services at low prices. If I understand DigitalOcean's current business model correctly, it somewhat resembles that of a street-level heroin dealer.
James
@vonskippy:
Ironically, fresh oats thru a horse - - not so good, yet fresh coffee beans thru a civet cat - - the most expensive coffee in the world.
+1
I use HostGator for some things, NFSN for something else, and Linode when I want full access to the server environment.
PS. I have been trying VPS6.net for a really cheap Windows server. It has actually been ok to play with for $7 per month. I still can't figure why the US Govt. refuses to use anything by MS after playing with MS servers for a while. Long live Gnu/Linux.
I just got this email in my inbox:
@DigitalOcean Support:
Even though we offer free unlimited bandwidth, we still have to monitor excessive bandwidth usuage. Your account has been flagged for network traffic and we will need to evaluate your account.
Please open a ticket to avoid any future bandwidth restrictions:
https://www.digitalocean.com/support We aplogize for any inconvenience and hope to deliver the best hosting experience possible.
I submitted a support ticket and this was the conversation that followed:
@Smark:
Hello,
I received an email in my inbox a moment ago notifying me that my bandwidth had been throttled. I would like to remind you that your website says in multiple places that your guys have unrestricted bandwidth usage:
Most importantly from your pricing page:
Q: Do you charge for bandwidth?
A: No. There are no charges for in-bound or out-bound bandwidth.
Q: Do you limit bandwidth?
A: No we do not limit monthly bandwidth in anyway. We allow you to fully utilize the network and burst traffic without limits.
And from your features page: FREE Bandwidth - We don’t charge for bandwidth. You’ll save a ton of money with our network and it's easy to get started. No need to figure out how much bandwidth you’re going to use, whether that traffic is in or out. Set your site up and pay no extra if it takes off. All servers come with 1Gb/sec. network interface.
To me this seems like false advertising. You explicitly state as one of your selling points that you allow us to FULLY UTILIZE THE NETWORK AND BURST TRAFFIC WITHOUT LIMITS. You state that you do not limit bandwidth in any way. It seems to me that throttling is limiting by another name. One could say that a speed LIMIT on a highway is likeable to thottling the speed of your car.
I had heard that Digital Ocean is providing excellent pricing with awesome flexibility. I thought it was "too good to be true" but I hoped I would be wrong.
Thank you for your time.
PS. Your bandwidth warning email has two misspellings in it.
@DigitalOcean Support:
Your server connection has been unthrottled. We try to maintain free unlimited bandwidth for our users. However, on occasion, due to abusive users on our system, legitimate users do get flagged.
I sincerely apologize for the inconvenience.
Thanks,
Etel
ps. we will get the bandwidth email corrected
@Smark:Hi Etel,
I would appreciate it if you could mark my account or at least my current droplet as white listed. I plan on utilizing a lot of bandwidth for legitimate purposes. I am abiding by your TOS/AUP. Not running a CDN or anything like that.
It would be unfortunate to have to submit a ticket weekly to get my server "unlocked".
Thank you for your time.
@DigitalOcean Support:Hi– I understand what you are looking to do. If you let us know let us know a little more about you and the traffic you plan to use, we can begin to whitelist your server:
Please let us know the following:
-Your Name
-Location
-Phone Number
-Type of Traffic you will be sending from the server
-1 or more of the following:
1. Your public twitter handle
2. Your blog
3. Your company or personal website
4. Your public facebook profile
Thanks,
Etel
@Smark:I'm not sure why you need those details. My name and other contact information can be found from information you already have.
I plan on using the my droplet as a webserver, a backup location for my personal files, and a VPN as I travel often and would like to have a secure connection.
I also do not use twitter, have a website, blog, or a facebook.
Thank you for your time and willingness to work with me!
@Smark:Upon further research I found the following:
https://www.digitalocean.com/community/ … -bandwidth">https://www.digitalocean.com/community/questions/do-you-limit-bandwidth It shows you stating (from the FAQ) what was discussed earlier. In addition someone asked about throttling:
Q: But do you throttle traffic, so that when it reaches a certain amount within a certain time period, it drops to a slower speed?
A: We do not throttle bandwidth or traffic after a certain amount. From time to time you may speed different bandwidth speeds depending on internet conditions, load on your droplet, or the system as a whole.
I believe this question was answered by your CEO, Ben Uretsky (DigitalOcean-Ben). It seems to me that the progress of this situation has gone as follows:
You say personally say in FAQ Bandwidth is free to use without limitations.
Someone specifically asks about bandwidth throttling
Your CEO states SPECIFICALLY that you do not throttle bandwidth.
I sign up for an account, pay real money for your services
I use your unlimited services
I get throttled for assumed abuse, which is an understandable issue
I request to be white listed to prevent myself from being throttled in the future (which shouldn't happen in the first place according to your CEO)
You imply that this is only possible if I give you information that was not necessary during initial sign up, including some odd requests such as my twitter account.
It seems to me you guys might be going back on your word just a bit. Please, PLEASE, assuage my suspicions.
Thank you for your time.
@DigitalOcean's Marketing Manager:We are always evaluating the services, products, and offers that we give out, and adjust accordingly.
Initially we offered free bandwidth because we didn't want to spend time writing bandwidth accounting code and wanted to focus on the core product of spinning up virtual servers.
We will most likely introduce bandwidth pricing in the future however we will not go back on our word because we are very grateful to all of our early customers that gave us a shot so anyone who registers while we are offering free bandwidth banner, we will continue to honor so long as it doesn't honestly cause any issues on the hypervisor for other users and abides by our TOS, which is the same way that we treat other resource like CPU or Disk I/O.
Thanks,
Moisey
@Smark:Moisey,
Thank you for your reply. I fully support your decision. I will hope that my server is white listed. If it causes an issue with performance on a host please let me know and we can take the situation from there. In the mean time until that point in time comes, I would greatly appreciate my server not being throttled.
Thank you for your time.
@vonskippy:
They lie, you lie (you state here you're running pirate bay junior, but to them you say it's a vpn gateway), where does it end?
I don't want to make this about me particularly, more about how they are a bit "too good to be true". I'm not running a "pirate bay junior", simply seeding for a private tracker. To be honest, as long as the content meets their AUP, it shouldn't really matter what I'm doing with it as long as it's not pointless.
I figure a VPN gateway is a close to a private SOCKS Proxy as I was willing to get. Proxy has a bad connotation, but it's effectively being used for the same reason here: I don't want my provider knowing what I'm doing. Whether that's Starbucks, or Cox Communications.
I'll update my post with further conversations between them, but I won't respond to any more posts in this thread, just wanted to share my experiences, not create any sort of drama.
Linode for life! Never once thought about changing.
@obs:
@sednet:Linode - Are you going to take this? How about lowering your costs for a while to keep these upstarts from stealing your customers?
If the "upstarts" raise their prices people will come back if they left, or more likely when something breaks and it takes ages to resolve instead of Linode's normal quick fix time then they'll come back.
Better for Linode if those people never leave at all.
Mostly I'm trying to hassle Caker for more stuff. Just repeat after me: We want more memory/bandwidth/disk capacity/disk speed/CPU and/or lower prices and/or cheap bulk storage and/or a linode 256 offering.
And I believe the current 512 was a 256 at one point. If not that, it was a 360 once.
@derfy:
My 512 is perfect for my usage. If you want / need more resources, upgrade.
And I believe the current 512 was a 256 at one point. If not that, it was a 360 once.
I want more for less money. There is a recession on you know.
I think the lowest end offering was an 80 running on UML not XEN some years back. A few people on here will remember running mailservers, spam filters, DNS servers, web servers, php, databases, and other things all in under 80 meg of ram. We were also limited by IO tokens so your disk speed went really slow if you swapped too much, or even just unpacked a big tarball. It was a nice challenge tuning all the stuff to work within those limits.
Right now I'd rather have half the capacity and half the price for some applications. I manage one machine that just runs BIND and tests connectivity to a few other IPs about once an hour. That would run very easily on 128Meg ram.
@sednet:
Mostly I'm trying to hassle Caker for more stuff.
Yes lets all be greedy bastards and see if we can't make Linode meet our cheap ass demands so that they too can be driven out of business.
Geesh people, if $20/month is too expensive, find a new hobby.
@vonskippy:
@sednet:Mostly I'm trying to hassle Caker for more stuff.
Yes lets all be greedy bastards and see if we can't make Linode meet our cheap ass demands so that they too can be driven out of business.Geesh people, if $20/month is too expensive, find a new hobby.
You don't know how many of those $20/month machines I manage. I don't just manage my own machines.
I negotiate with suppliers for a better deal as does anyone who isn't just burning money for a hobby. Only a fool tells a supplier that their price is OK or what they get for that price is enough.
@sednet:
Only a fool tells a supplier that their price is OK or what they get for that price is enough.
Every time you pay your bill you are telling your supplier exactly that. Linode has chosen a market position that doesn't attempt to compete on price with lower-cost providers, and the fact that you continue to manage a multitude of linodes indicates that in the big picture you don't really want them to become more like those lower-cost options.
@Stever:
@sednet:Only a fool tells a supplier that their price is OK or what they get for that price is enough.
Every time you pay your bill you are telling your supplier exactly that. Linode has chosen a market position that doesn't attempt to compete on price with lower-cost providers, and the fact that you continue to manage a multitude of linodes indicates that in the big picture you don't really want them to become more like those lower-cost options.
Linode does compete on price on the low end stuff very nicely. However, there is always room for improvement. The IT world keeps offering bigger and faster for less money. If Caker actually listened to the fanboys on here nothing would ever improve because it's already perfect, Linode would stagnate and die. That's not what has happened. Instead Linode grew from a tiny little company to a huge one because it kept offering more and more. As a customer this should be encouraged.
Paying a bill doesn't mean you are entirely happy. It doesn't mean you would not jump ship as soon as a better offer comes along. It doesn't mean you should not gently pressure your supplier to give you more.
If you're ordering a very large number of linodes, they may be willing to negotiate a volume discount. I've no idea if they do such a thing, but if you've got hundreds of linodes, you probably represent a big enough customer to negotiate that sort of thing.
@deadwalrus:
![](" /> You are meant to put a caption on it to make some kind of funny point.
@Digital Ocean Support:
During the above maintenance window we will be doing work on our network in New York. We will need to reconfigure our network settings which may cause a brief network outage.
Our maintenance window may affect your servers: '[redacted]'
Maintenance Window Start Time: Tuesday January 22nd 22:00 EST
Maintenance Window End Time: Tuesday January 22nd 24:00 EST
Maintenance Window Duration: The reboot should not take longer than 15 minutes.
Impact: There will be a loss of connectivity as the core routers reboot.
Reason: Onsite engineers will reboot the core routers as part of a planned maintenance.
Resolution: We will be monitoring all servers & services closely, if you experience any interruption longer than 15 minutes please contact us to investigate further.
https://www.digitalocean.com/support
I wasn't aware that 10 hours notice for 15min downtime was acceptable. They have onsite engineers rebooting core routers as part of planned maintenance?
@Guspaz:
I wasn't aware that changing network settings required rebooting enterprise-grade routers.
Apparently they need time to polish their plugs.
@derfy:
My 512 is perfect for my usage. If you want / need more resources, upgrade.
And I believe the current 512 was a 256 at one point. If not that, it was a 360 once.
It was a 64…
From my yearly billing email…
Linode 64 & 1.5 GB Disk FREE 05/01/2004 - 04/30/2005 $0.00 $239.40
Linode 64 & 1.5 GB Disk FREE 05/01/2005 - 04/30/2006 $0.00 $239.40
Linode 80 & 1.5 GB Disk FREE 05/01/2006 - 04/30/2007 $0.00 $239.40
Linode 256 & 4 GB Disk FREE 05/01/2007 - 04/30/2008 $0.00 $239.40
Linode 360 & 5 GB Disk FREE 05/01/2008 - 04/30/2009 $0.00 $239.40
Linode 360 & 8 GB Disk FRE05/01/2009 - 04/30/2010 $0.00 $239.40
Linode 360 & 8 GB Disk FRE05/01/2010 - 04/30/2011 $0.00 $239.40
Linode 512 & 8 GB Disk FRE05/01/2011 - 04/30/2012 $0.00 $239.40
Linode 512 & 10 GB Disk FR05/01/2012 - 04/30/2013 $0.00 $239.40
I'm gonna guess this year will also show as "512 & 10 GB Disk FR" because the extra disk space this year didn't propagate into grandfathered allowances.
You can tell my system was originally designed for a 64Mb machine…
total used free shared buffers cached
Mem: 496 442 53 0 9 369
-/+ buffers/cache: 63 433
Swap: 256 6 250
@IRC #digitalocean @ Freenode:
15:22 <@clone1018> Like popular, so we could have 64 cores (with ht) machines
15:22 < Kyle> wait a second, planned downtime of 15 minutes.
15:23 < SSL> in NY only
15:23 < SSL> i think
15:23 <@clone1018> Yes
15:23 < Kyle> clone1018: that breaks the SLA, no?
15:24 < Kyle> isn't the SLA 99.99 percent?
15:24 * Kyle looks
15:24 <@clone1018> It depends, is our SLA based on year or month?
15:24 < Fishfish0001> month
15:24 < Fishfish0001> I believe.
15:24 < Kyle> 446.4 is your max downtime allowed per month
15:24 <@clone1018> Why do you say that?
15:24 < Kyle> in seconds
15:24 <@clone1018> Kyle: seconds?
15:25 < SSL> SLA is only for unexpected downtime
15:25 < SSL> that one is expected
15:25 < SSL> like
15:25 < SSL> announced
15:25 < Kyle> and since when does configuration changes require reboots on enterprise routers?
15:26 < mwcampbell> Kyle: Even I am not that anal. If I really care about uptime, I'll pay for geographic redundancy.
15:26 < Fishfish0001> 43 minutes is .1% of 99.9% of 30 days in minutes
15:26 < Kyle> Fishfish0001: it's 99.99
;) 15:26 < Fishfish0001> oh nvm
15:27 <+KamalN> .tell asdasdads no it's not me
15:27 < Fishfish0001> 4 minutes, 19 seconds
15:27 * Kyle shrugs
15:27 < Kyle> I'm just pointing it out :p
15:28 -!- oDiafanos [~
odifis@athedsl-247412.home.otenet.gr ] has joined #digitalocean15:28 < Kyle> Also, as far as the site mentions, there is no definition of "Uptime" beyond uptime,
15:28 < Kyle> so, not sure where you got that SSL
15:29 <@vt0r> Kyle, SLA's do not apply for scheduled maintenance
15:29 < Kyle> vt0r: then, 15:25 < Kyle> and since when does configuration changes require reboots on enterprise routers?
15:29 < SSL> that's what i said
15:29 <@clone1018> Kyle: It's probably more advanced then that
15:29 <@vt0r> Who told you it was just config changes?
15:30 < Kyle> vt0r: then clarify it :p
15:30 <@clone1018> But explaining every process they'll be going through is time consuming and unnecessary.
15:30 <@clone1018> Especially since it's 15 minutes, go watch an episode of scrubs and laugh at your downtime emails
15:31 < Fishfish0001> The downtime was supposed to happen last year
:( 15:32 < Fishfish0001>
http://i.imgur.com/3ylZgxr.jpg 15:32 <@clone1018> .wa time until Tuesday January 22nd 22:00 EST
15:32 < MrSwimmy> clone1018:
| Time span: 6 hours 27 minutes 25 seconds 0.269 days 6.457 hours 387.4 minutes 23246 seconds http://www.wolframalpha.com/input/?i=ti … 2%3A00+EST">http://www.wolframalpha.com/input/?i=time+until+Tuesday+January+22nd+22%3A00+EST
From #linode:
@IRC #linode @ OFTC:
15:39 < Nightmare> Oh my god
15:39 < Nightmare> I know clone1018 personally >.>
15:40 < KyleXY> Nightmare: You should tell him he just fucked up, badly.
15:40 < tubaguy50035> Nightmare: well. He's had a good run. Time for him to die.
15:40 < Nightmare> KyleXY: He used to run a Minecraft hosting company. It died off rather quickly.
Snips from yesterday:
@#digitalocean @ Freenode:
14:56 < Adran> How do you have a one gigabit pipe per vps?
14:58 <+KamalN> Why not?
14:59 <@clone1018> Adran: You grab the 1Gbit cable
14:59 <@clone1018> well
14:59 <@clone1018> actually sorry
14:59 < Adran> lol
14:59 <+KamalN> clone1018: 10Gbit
14:59 <@clone1018> You grab two 10Gbit pipes
14:59 <@clone1018> Throw them into the machine
14:59 <@clone1018> and boom
14:59 <@clone1018> Allocate 1Gbit per droplet and done!
14:59 <@clone1018> (Please note I have no idea how DO does it so ignore me
:D )15:00 < Adran> aren't you employed by digital ocean?
15:00 <@clone1018> Yes but I don't work in the server room
15:00 <@clone1018> I help them remotely with support
15:00 < Adran> ah
15:00 <@clone1018> I also work at an eyeglasses company but I couldn't tell you what software we use for shipping
:D 15:06 < Tux> o.o
15:06 < Tux> 1gbit per vps.
15:09 <@vt0r> defunctzombie, "without swap, if you build anything or something peaks memory for a moment it will just crash"
15:09 <@vt0r> The same thing will happen if you exceed swap capacity
15:10 <@vt0r> There's no reason to allow overutilization of RAM, because disk virtual memory is very slow
15:10 <@clone1018> Even SSD based virtual memory
An SLA guarantees a minimum of downtime. It shouldn't matter if it's scheduled or not, either way it's downtime.
Jeff
@obs:
I think that network maintenance may have screwed some things up. I created a new server just to burn my free credit with some tests and as soon as it was created it wasn't reachable from the internet, I raise a support ticket, they said they'd look into it, that was 6 hours ago. I give up I can't even use my free credit!
Wonderful first impression, no?
21:26:04 -!- jooajsd [d1008f71@gateway/web/freenode/ip.209.0.143.113] has joined #digitalocean
21:26:20 < jooajsd> when was this company foudned
21:34:57 -!- jlott24 [~james@wifi02.la3.routers.zerolag.com] has quit [Ping timeout: 276 seconds]
21:35:15 < Fixxar> DO?
21:35:30 < Fixxar> i would guess 2011 or so
21:35:40 < Fixxar> but they have been in the hosting business for over 10
This piqued my curiosity, as I had a friend who was with clone's old hosting company, so I searched through my logs to see if I could find anything interesting relating to that. Something I found worth noting are logs of clone1018 threatening to falsely send a cease and desist to a user's ISP to get him off the internet purely because he didn't like him. The logs cited a pastebin link, which seems to be still up at
EsperNet-#risucraft | Apr 10 18:41:53 <clone1018> lahwran: operation ban langkid1 from the internet is a go
EsperNet-#risucraft | Apr 10 18:42:40 <clone1018> InsanityBringer: just send his parents a C&D for downloading torrents or something
EsperNet-#risucraft | Apr 10 18:42:50 <clone1018> they'll ground him right away</clone1018></clone1018></clone1018>
A search through all my logs for anything containing 'clone1018' showed a few other interesting tidbits. For example, the very first line that came up in the output log was:
EsperNet-langkid1 | Jan 30 22:12:28 <langkid1> well, my VPS is down 'cuz of clone1018's maintenace</langkid1>
I didn't look very thoroughly, but there were more than a few bits where he acted fairly arrogantly (in my personal opinion) or was rude to others. See:
EsperNet-#bukkit | Feb 09 10:22:13 <clone1018>no shit
EsperNet-#bukkit | Feb 09 10:22:55 <clone1018>I'm making sure I can cater to idiots
EsperNet-#bukkit | Feb 09 10:26:09 <clone1018>What the hell is stopping fire on my server
EsperNet-#bukkit | Feb 09 10:26:10 <clone1018>agagagdhdhdd
EsperNet-#bukkit | Feb 09 10:26:20 <clone1018>Nope
EsperNet-#bukkit | Feb 09 10:26:35 <clone1018>wtf who uses essentials
EsperNet-#bukkit | Feb 09 10:27:11 <clone1018>BananaChunk.jar ChatManager.jar CraftIRC.jar FalseBookCart.jar FalseBookExtra.jar Modifyworld.jar PermissionsEx.jar Register-1.5.jar WorldEdit.jar dynmap.jar nChat-0.7.jar Biome.jar CommandBook.jar FalseBookBlock.jar FalseBookCore.jar HawkEye.jar Permissions.jar Questioner.jar Towny.jar WorldGuard.jar mcMMO.jar
EsperNet-#bukkit | Feb 09 10:27:13 <clone1018>those are the plugins
EsperNet-#bukkit | Feb 09 10:28:20 <@TnT> clone1018: I can't see anything other than WG that would do it...
EsperNet-#bukkit | Feb 09 10:28:29 <clone1018>I think Towny might do it
EsperNet-#bukkit | Feb 09 10:29:08 <jbdye>clone1018: i use essentials
EsperNet-#bukkit | Feb 09 10:29:24 <clone1018>CommandBook?
EsperNet-#bukkit | Feb 09 10:29:45 <clone1018>Then make command aliases?
EsperNet-#bukkit | Feb 09 10:30:07 <clone1018>Wait
EsperNet-#bukkit | Feb 09 10:30:10 <clone1018>TnT: who the fuck are you
EsperNet-#bukkit | Feb 09 10:30:16 <clone1018>why are there so many ops
EsperNet-#bukkit | Dec 23 19:59:45 <clone1018>CosmicVoyager: DONT FUCKING SPAM
EsperNet-#bukkit | Dec 23 19:59:50 <clone1018>JESUS CHRIST USE A FUCKING PASTEBIN</clone1018></clone1018></clone1018></clone1018></clone1018></clone1018></clone1018></jbdye></clone1018></clone1018></clone1018></clone1018></clone1018></clone1018></clone1018></clone1018></clone1018>
I don't know about you guys, but I'd feel very wary about giving this guy money for anything.
@Guspaz:
Scheduled downtime should absolutely count against an SLA.
An SLA guarantees a minimum of downtime. It shouldn't matter if it's scheduled or not, either way it's downtime.
You have to read the SLA. Some do count scheduled downtime as downtime, some don't.
A place I worked for once specified a 98% operating system uptime in their SLA's. This was just for servers internal to that company but even so it was amazingly lame. We could have turned the servers off for a week every year and still met that.
@DigitalOcean:
YOU'VE BEEN GRANDFATHERED IN!
We're so grateful for all of our early adopters and wanted to say thanks by grandfathering your account so you will receive free bandwidth FOREVER!
For everyone else, DigitalOcean will be introducing bandwidth pricing. Plans start with 1TB of bandwidth and will increase incrementally. Once the monthly transfer limit has been exceeded, it is only 2 Cents per GB thereafter.
Click here to login to your account
Their website also has been updated to not include unlimited bandwidth.
I have spent many years working within hosting and alongside the success I had, I've also made mistakes, mistakes made due to being over enthusiastic and over extending myself, mistakes that I very much regret. DigitalOcean is a fantastic company who has given me an opportunity to provide value to the industry I care most about. I've learned from the mistakes I made and they won't be made again, especially not at DigitalOcean where I'm a "Part Time Support Member".
Thanks for your time.
Nightmare, if you want to talk to me privately about any issues you've had with me in the past you can reach me at
@vonskippy:
But….but…. it's cheap and with UNLIMITED bandwidth - that means it rocks right?
Not anymore!
As it is, it would appear to have turned-out very well indeed for Linode - not so good for the competition
@Mr Nod:
Credit to Linode themselves for allowing a discussion about a competitor in their public forums.
As it is, it would appear to have turned-out very well indeed for Linode - not so good for the competition
:D @NeonNero:Agreed. Reading through this thread about the experiences people have had with this competitor just gives me one extra reason to stick with Linode.
:-)
Because really, most of these "competitors" that come into Linode's general forums are usually crap, and they don't compare to Linode standards whatsoever. So why worry? If they lose a few customers, chances are they'll come back when crap hits the fan.
In this case, they didn't have any legitimate reason why they could offer what they promised for the prices they did, so it turned out that they were offering prices that low through lies and terrible service.
@Cromulent:
I've just spun up a 512MB instance on their free 7 day trial and the speed is awful. My Linode 512MB beats the hell out of it in both network performance and hard drive performance.
Still, since I'm planning on using it as an email server I guess speed is not crucial, but if I was doing something to be important I'd stick with Linode.
Cromulent – what benchmarks did you use to measure performance? I ran some benchmarks of CPU, RAM, disk, and bandwidth, and found that Linode won only on the bandwidth tests, with a comparable Digital Ocean node performing significantly better on all other benchmarks I tried.
I did the bandwidth tests with tespeedPhoronix test suite
~~http://i.imgur.com/01ZG5lC.png
It's not a 100% apples-to-apples comparison since my Linode is serving a couple of very low-volume web sites, but I didn't see any traffic at all coming in during these tests, and there was plenty of spare CPU/RAM headroom.
Don't get me wrong – I've been with Linode for 4.5 years, and I'm deinitely not here to bash the service, but I would like to see what benchmarks other folks are getting on their nodes to see if my results are atypical. My node is a Linode 768 instance, and I've felt that it's been quite sluggish for a long time now, which is what led me to look into Digital Ocean.
Perhaps I'm just on a crowded host, but for $30/month, I'm feeling like the results on my Linode should be better, and Digital Ocean is looking pretty good, performance-wise. I agree with everyone who's said that DO is probably dumping their plans at a loss to gain market share, and will probably end up in the "oversold" category once this catches up to them, but if I can get six months out of them, I can save $150, and that's not chump change. I doubt the service will measure up to what Linode has given, but I just can't pay a 400% premium for slightly support response time and more honesty about bandwidth numbers. Perhaps if the performance gap was 2:1 instead of 3:1 or more I could do it, but as things are, I'm going to at least give DO a shot for a month or two to see how reliable it is, and if it stays solid, I'll definitely be considering a switch.
EDIT: I originally just ran the disk write numbers, but upon further review, Linode does significantly better on reads. I updated the charts with a breakdown of sequential/random reads/writes.~~
I've been with probably 7 or 8 since the late 90's. Linode not only is the best performing virtual machine service, it feels like running on a real machine. I don't need a set of statistics to tell me that.
Perhaps I'm just on a crowded host
From my experience, that doesn't happen here. Perhaps it's something else.
@jebblue:
it feels like running on a real machine. I don't need a set of statistics to tell me that.
With all due respect, that's just silly. Benchmarks aren't gospel, but they're certainly better than a subjective "feels like" analysis, and in any event, my Linode feels slower than a comparable DO node. Your experience may be different, but that doesn't make my experience any less valid, especially given that I have numbers showing many metrics where my Linode compares unfavorably.
@jebblue:
From my experience, that doesn't happen here. Perhaps it's something else.
Well, it happened once to me, and performance improved significantly after the Linode staff migrated my node. Perhaps it was something fluky about the host I was on, but your experience is different than mine. Again, I've been very pleased overall with my Linode service over the years, but I don't think it's helpful to be so dismissive of someone else's experience because yours happens to be different.
I'm willing to consider that there might be something else going on with this host, so my plan now is to migrate my services to my DO node and then configure a new Linode from scratch to repeat the benchmarks. I'll post those results here once I have them.
@Guspaz:
Your sequential write benchmark doesn't make sense… It's basically zero for linode, which would indicate a problem with your benchmark.
It was actually 6.4 MB/s, not zero, but compared to the 300+ it didn't show up well on the graph.
Like I said, I will try to re-run the benchmarks from a fresh OS install with nothing hitting my server to see if things improve, but I do think the benchmarks reflect what's actually happening on my Linode.
most of my content gets generated periodically and cached on disk and in memory and read many times before being deleted and regenerated.
also benchmarks don't test the reliability of a services, or the support.
DO are using unmirrored or JBOD SSDs.
Just guessing, I could be totally wrong.
@sednet:
My guess is that Linode is using Raid 1 on 15K disks thus doubling read speed and halving write speed due to not using write behind caching, or the cache being full.
DO are using unmirrored or JBOD SSDs.
Just guessing, I could be totally wrong.
RAID 10, I believe.
More interesting benchmarks would be AB and MYSQL Benchmark.
Throughput to the real world is all I care about - don't care which box has faster internal tests (at least for remote VPS's).
I can't imagine what would get me to consider leaving Linode. Well, maybe if caker sold it and bought an island or something
Instead of people digging them, why not just see them as a cheap alternative offering something different. Who knows, they might grow, they might convert their good wishes to a good business - wish them luck and go on your way
EDIT: Actually I might throw my (low bandwidth) sites onto one of them.
We can only hope he and his team would start another one.
tl:dr DigitalOcean has better performance per dollar, generally speaking. Linode has the edge when it comes to CPU which will be more apparent when you do something like transcode a video.
I'm currently developing my SaaS offering and evaluating VPS providers. My testing environment is on Linode and I'm very happy with what I've seen. My current plan is go with Linode to host all my clients' boxes.
However, for the same price as a Linode 512 I can run 2 x 1GB nodes at DigitalOcean. Apparently local LAN is coming in two months time so I can actually take advantage of them. More good reports and benchmarks are appearing about DO. Twitter commentary is positive.
I want to stick with Linode. I like the idea of basing my entire business and livelihood on a mature company with proven credentials. I don't want to rebuild my server images all over again. But at the moment DigitalOcean have just got Linode crushed on a performance per dollar basis. And I mean crushed so badly that I almost have no choice, commercial pressures being as they are.
I'll be watching eagerly to see how Linode respond. For me to stay with them they don't even have to match Digital Ocean's pricing. They just need to meet them halfway (get in the same ballpark, even) and let their brand/track record take care of the rest.
@Guspaz:
Well, when your services are priced at what are obviously a steep loss, FUD is bound to follow. It turned out to be largely justified too, because when they got called out on how silly their plans were, they put caps on the bandwidth usage.
I looked into that allegation and they claim to be profitable. They have also stated that prices will not rise. That said, I do share your concerns that the pricing will result in overselling. I also value track-record very highly, but not necessarily 300% more highly.
People spread the same FUD about Cloudflare when they upset the traditional CDN business model. But they're going from strength to strength and their service is still free for Mom & Pops, despite what the doom-mongers forecast. We could have a similar scenario here.
@TeshooLama:
I want to stick with Linode. I like the idea of basing my entire business and livelihood on a mature company with proven credentials. I don't want to rebuild my server images all over again.
Wait a sec. You don't need to rebuild your server images to copy them from one host to another. Boot off a recovery OS and copy the disk images anywhere you like.
I used to install OpenBSD on dedicated servers like that.
@TeshooLama:
@Guspaz:Well, when your services are priced at what are obviously a steep loss, FUD is bound to follow. It turned out to be largely justified too, because when they got called out on how silly their plans were, they put caps on the bandwidth usage.
I looked into that allegation and they claim to be profitable. They have also stated that prices will not rise. That said, I do share your concerns that the pricing will result in overselling. I also value track-record very highly, but not necessarily 300% more highly.
People spread the same FUD about Cloudflare when they upset the traditional CDN business model. But they're going from strength to strength and their service is still free for Mom & Pops, despite what the doom-mongers forecast. We could have a similar scenario here.
They also claimed that unlimited bandwidth was sustainable at the $5 pricepoint, but when people actually tried to use it, they got shut down and the unlimited disappeared. Considering that the one-time cost for 20GB of enterprise-grade SSD storage is probably $50-60 on top of all the other costs, how can they be profitable at $5/mth?
@Guspaz:
@TeshooLama:
@Guspaz:Well, when your services are priced at what are obviously a steep loss, FUD is bound to follow. It turned out to be largely justified too, because when they got called out on how silly their plans were, they put caps on the bandwidth usage.
I looked into that allegation and they claim to be profitable. They have also stated that prices will not rise. That said, I do share your concerns that the pricing will result in overselling. I also value track-record very highly, but not necessarily 300% more highly.
People spread the same FUD about Cloudflare when they upset the traditional CDN business model. But they're going from strength to strength and their service is still free for Mom & Pops, despite what the doom-mongers forecast. We could have a similar scenario here.
They also claimed that unlimited bandwidth was sustainable at the $5 pricepoint, but when people actually tried to use it, they got shut down and the unlimited disappeared. Considering that the one-time cost for 20GB of enterprise-grade SSD storage is probably $50-60 on top of all the other costs, how can they be profitable at $5/mth?
Well, like you said. It's a one-off cost. Sweat those assets for 5 years and you have a positive ROI.
Thanks for the tip Sednet. I will try that.
@TeshooLama:
Sweat those assets for 5 years and you have a positive ROI.
Lol - you're joking right?
Any business model that has a 5 year lead time before profits is doomed - especially in the IT service industry.
Besides, when have you EVER had server hardware running 24/7 last 5 years?
@TeshooLama:
@Guspaz:They also claimed that unlimited bandwidth was sustainable at the $5 pricepoint, but when people actually tried to use it, they got shut down and the unlimited disappeared. Considering that the one-time cost for 20GB of enterprise-grade SSD storage is probably $50-60 on top of all the other costs, how can they be profitable at $5/mth?
Well, like you said. It's a one-off cost. Sweat those assets for 5 years and you have a positive ROI.
Thanks for the tip Sednet. I will try that.
Okay, $50-60 might not sound like much. But then think about - cost of other server parts. You'll probably also want at least 2, preferably 4 drives in a server in RAID. You'll want manpower, etc., etc.
We haven't even started to figure in costs of sending the parts to repair (even if you have a guarantee, you'll probably still pay for shipping), keeping spare parts, server colo fees, storage for those spare parts (they need to be somewhere in the DC you're coloed in for easy access, 24/7 staff in the DC.
Do you still think they could survive on $5/mo. realistically?
@vonskippy:
@TeshooLama:Sweat those assets for 5 years and you have a positive ROI.
Lol - you're joking right?Any business model that has a 5 year lead time before profits is doomed - especially in the IT service industry.
Besides, when have you EVER had server hardware running 24/7 last 5 years?
I should've said 'nice ROI'. It might be cashflow positive in a year for all we know. The bulk of profits might be derived from the top 20% of customers and the rest might be run at breakeven for all we know.
What do we really know? All I'm hearing are the same accusations that were levelled at Cloudflare, which turned out to be complete garbage.
@TeshooLama:
@vonskippy:
@TeshooLama:Sweat those assets for 5 years and you have a positive ROI.
Lol - you're joking right?Any business model that has a 5 year lead time before profits is doomed - especially in the IT service industry.
Besides, when have you EVER had server hardware running 24/7 last 5 years?
I should've said 'nice ROI'. It might be cashflow positive in a year for all we know. The bulk of profits might be derived from the top 20% of customers and the rest might be run at breakeven for all we know.
What do we really know? All I'm hearing are the same accusations that were levelled at Cloudflare, which turned out to be complete garbage.
I hate to say it, but I am tending to agree with you. I'm not quite ready to move my production system over – I'll wait a bit longer to see if they prove reliable. But for f***s sake people -- the benchmarks show them beating Linode and they are 1/3 the price. Why so defensive?
The reason people are hating on DigitalOcean is because of:
If it sounds too good to be true, it probably is.
There's no such thing as a free (or unreasonably cheap) lunch.
You get what you pay for.
When somebody is selling you a hosting service that they can't possibly break even on for 5 years, there's something funky going on.
@Guspaz:
Errm, the benchmarks I saw put them well behind Linode in everything but disk IO…
The reason people are hating on DigitalOcean is because of:
If it sounds too good to be true, it probably is.
There's no such thing as a free (or unreasonably cheap) lunch.
You get what you pay for.
When somebody is selling you a hosting service that they can't possibly break even on for 5 years, there's something funky going on.
Do you think their VCs will be that naive? It's more likely their business plan is to run the bottom tier at breakeven to build the brand and market share, while looking to the middle and upper tier products to bring in the dough.
Shared web hosting was expensive in the 90s. Now it costs $1 per month.
@Guspaz:
Errm, the benchmarks I saw put them well behind Linode in everything but disk IO…
The reason people are hating on DigitalOcean is because of:
If it sounds too good to be true, it probably is.
There's no such thing as a free (or unreasonably cheap) lunch.
You get what you pay for.
When somebody is selling you a hosting service that they can't possibly break even on for 5 years, there's something funky going on.
I once heard about a company giving away 5 gigs of online persistent, redundant, fast storage for free, along with their GUI interfaces to dozens of operating systems. It was unheard of at the time. They also came out of the same startup incubator as digital ocean and are backed by many of the very same VCs. Obviously it was too good to be true and couldn't last. You just can't undercut the established players like that.
Now this little fly-by-night company called DropBox is valued at $5 billion. Total scam though – no way they can pay for their costs by giving away that bandwidth and storage for free. Any day now I know they will go the way of the dodo. Only a matter of time. Loss leaders always fail and go out of business.
deduplication to reduce storage requirements
That's one way to make a profit.
Also, the description of dropbox is exaggerated. They only ever offered two gigs free by default (deduplicated, as Hoopy pointed out), rather than five, it was never particularly fast, they've never supported dozens of operating systems (even including mobile platforms they currently support only six). They also priced their non-free tiers reasonably, making the free tier a loss-leader with minimal cost to them. DigitalOcean, on the other hand, has questionable prices for all pricepoints, although I can see how margins would be larger the higher you go. Linode has shown, however, that the vast majority of customers are going to go for the lowest few packages.
Ok - so 2 gigs instead of 5; 6 OSs instead of 12. I don't see how this impacts my point.
I'm still wondering about your defensiveness here. I understand fully that Linode provides a great service; I have been with them for many years and have always been satisfied. I have simply noticed – as apparently have many others -- that Linode's price/performance is starting to get noticeably out-of-whack with the up-and-comers. You yourself have acknowledged that we haven't seen legitimate hardware upgrades for about 2 years.
Regardless, the most salient point to me is that I haven't really seen anything wrong with DigitalOcean (other than promising unlimited bandwidth then quickly retreating). You can drum up as much fear as you want (and you seem to enjoy it), but the simple truth is that I can move a disk image over there, for minimal cost and very little risk, and save a lot of money and get similar/better performance. Yes, they might fold up -- and so might Linode -- but from what I see, their financial backers and technical advisers appear top notch.
The bottom line is: It seems entirely reasonable to me to ask why, when hardware prices decrease exponentially year-after-year, is Linode falling behind?
For what it's worth, I was with Slicehost for years before they got greedy, their prices drifted ever behind, and they got overtaken by an upstart called Linode. There were just as many -- if not more -- fanboys over there defending them. Is it so impossible to believe Linode could eventually head in that direction?
too good to be true?
I have been using Dreamhost ($10/month) for over 8 years. Still with them. Ten days back I sign-ed up for Linode-512 and I am quite impressed. I don't mind paying 20 bucks a month for VPS –- 24GB disk and 1TB transfer meets my needs but I prefer 1GB of RAM. Then I heard about DigitalOcean and signed up for it yesterday to check them out. While not as polished as Linode --- which is to be expected --- I found it quite good. I could do a side-by-side comparison of running nginx, php-fpm, https etc and the 1 core of DigitalOcean was not limiting. I have the option of trying the 1GB and 2GB plans (2 cores), pay for what I use, and still pay the same or less than what I would pay for Linode.
So, what is the problem? The problem is that it is too good to be true. DigitalOcean must be losing money at these prices and what happens when the seed money runs out say in 6 or 8 months? $1.25 per week wouldn't probably pay even the power bill. Lots of money will be required to expand beyond NY and Amsterdam. Linode would typically give free upgrades every June and if they upgrade the RAM to 1 GB and make backups free then it would make Linode look quite good. Linode has been around for 8 years, they may charge more but the probability of them being around a year or more from now is almost 100%. My worry is that if DigitalOcean folds without warning a year from now and doesn't give option to copy your data it would be a problem. So we have to have frequent backup outside of DigitalOcean to account for this.
Linode is putting forty 512M nodes on a server. Is DigitalOcean putting 160 nodes?
Babu
Tweet
6 Replies
We're going to be celebrating our two year anniversary this coming June so while we haven't been around for 8 years, we have been at this for a bit of time.
While I can't comment on the specifics of our business I can tell you that we are profitable and we are looking to provide the best service possible to our customers and also to continuously improve on the product.
Aside from that I would say sign up, use the product, and let us know how we can improve.
Thanks!
Answered by raiyu
February 24th, 2013 06:12
[
Also note, that this is the same company that failed miserably before because an ex-employee got upset and started deleting everything.
I try not to be a Linode fanboy, and I was very interested in DO because the price is so attractive. I don't care if they're profitable or not, I bet 1&1 is profitable too. There's something to be said about a company that treats its customer so well like Linode does.
Yes, we all wish they'd upgrade more stuff. The only issue I have with Linode as of late is that the VM's aren't on a SAN. There are no live migrations like you have at DO. This also limits Linode's hard drive capacity.
Again, to me reliability and a company that has enough integrity to let us argue about their competitors on their forum is worth the extra money.
Does that mean Linode should just sit back and enjoy its current position? No. Competition is good for consumers. I hope that DO makes Linode an even better company and pushes them beyond what they thought they could do.
All this talk about them is quite dumb. If you want to go there because of the price, go right ahead. You have limited features (cough cough no IPv6 or IP failover) and you have tons of people who run "questionable" things that get DDoS'd constantly. If you want to go, just go. Why stay here trying to get us to convince you? I think that's the "fighting for Linode" you're getting. I don't think anyone understands why you're still here if you really want to move to DO so bad.
You make a few salient points - but for FSM's sake, learn how to freaking post.
Is it "that" hard to figure out the quoting system on this forum?
I'm just a scientist not a code monkey (read today's The Daily WTF
Proof your posts so you don't look like a dumbass.
@deadwalrus
You make a few salient points - but for FSM's sake, learn how to freaking post.
Is it "that" hard to figure out the quoting system on this forum?
I'm just a scientist not a code monkey (read today's The Daily WTF - they rip PhD's pretty good in today's comments) and yet I manage to figure out how to quote a previous post (I've even mastered nested quotes).
Proof your posts so you don't look like a dumbass.
Ok wise one, teach me. I simply hit the "Quote" button and that's what happens. If that doesn't work – which it apparently does not -- I am not going to spend time digging through the help to figure that out, because quite frankly, I don't care.
During the time that I tried their service, I had hours of downtime. So much so that it got to the point of them just giving me my whole month free. Granted that's only $5, but come on? HOURS OF DOWN TIME?
Also note, that this is the same company that failed miserably before because an ex-employee got upset and started deleting everything.
This is the sort of argument that is convincing. I am, like I said, quite pleased with Linode, and have been for years. Whatever minor annoyances I may have with seeing competitors beat them on price is adequately addressed by posts like this, which point out first-hand experience of down times. I am willing to pay a premium, as I expect most are, to avoid that sort of issue.
My prior comments were directed to the non-first-hand experience FUD that I saw getting thrown around. Not convincing and not productive, and comes off as rampant Linode fanboyism.
@deadwalrus:
Ok wise one, teach me.
Since the tags look to be in order, I suspect you probably have 'Disable BBcode' checked.
@deadwalrus:
@tubaguy50035:During the time that I tried their service, I had hours of downtime. So much so that it got to the point of them just giving me my whole month free. Granted that's only $5, but come on? HOURS OF DOWN TIME?
Also note, that this is the same company that failed miserably before because an ex-employee got upset and started deleting everything.
This is the sort of argument that is convincing. I am, like I said, quite pleased with Linode, and have been for years. Whatever minor annoyances I may have with seeing competitors beat them on price is adequately addressed by posts like this, which point out first-hand experience of down times. I am willing to pay a premium, as I expect most are, to avoid that sort of issue.
My prior comments were directed to the non-first-hand experience FUD that I saw getting thrown around. Not convincing and not productive, and comes off as rampant Linode fanboyism.
I'm glad I was able to help. I encourage you to try it out yourself. We can piss and moan in this thread all day, or we can actually try things out and see if they fit our needs. DO will give you a free trial credit if you ask for it. I encourage you to have another box pinging the DO box and watch it go down for hours.
@Guspaz:
If they've been around for two years, why do they still seem like a beta service? The first time anybody actually tried to use significant amounts of bandwidth on the unlimited plans, they got rid of unlimited plans. Does that mean that for two years, no customer ever tried to use a substantial amount of bandwidth?
Good point. Although to be fair, he was running a seedbox to try to max out bandwidth. I agree they should not be advertising it if they don't intend to allow people to do it.
See it wasn't rocket science.
@vonskippy:
^^nicely formatted^^
See it wasn't rocket science.
Nope. Not sure why "Disable BBCode" is always checked for me by default – I can't imagine why anyone would want to do that.
Oh well.
@deadwalrus:
@vonskippy:^^nicely formatted^^
See it wasn't rocket science.
Nope. Not sure why "Disable BBCode" is always checked for me by default – I can't imagine why anyone would want to do that.
Oh well.
In User Control Panel:
Then I'm going to write a poorly formatted post about it.
Take that vonskippy!
@Nightmare:
Good luck with that, Derp Ocean doesn't offer IPv6
;)
I could always setup a SIXXS tunnel.
@sednet:
I'm going to setup a DIGITALOCEAN machine with CPANEL so I can use its superior disk IO with SYSTEMD to send BITCOINS over IPV6.
[kowtow] We're not worthy! We're not worthy! [kowtow]
James, impressed and laughing
@deadwalrus:
You yourself have acknowledged that we haven't seen legitimate hardware upgrades for about 2 years.
I guess I must have been imagining that 20% disk increase a mere 2 months ago. Ah well…
> It seems entirely reasonable to me to ask why, when hardware prices decrease exponentially year-after-year, is Linode falling behind?
Decrease exponentially? Wow, umm. No. If you drew a graph of that then people would be paying you to take their computers.
You're a funny man…
@sweh:
@deadwalrus:You yourself have acknowledged that we haven't seen legitimate hardware upgrades for about 2 years.
This was discussed before. When was the hard drive upgrade before that? 20% storage increase over 4 or 5 years is not exactly something to stand up and shout about. See hard drive cost graph below.
>
> It seems entirely reasonable to me to ask why, when hardware prices decrease exponentially year-after-year, is Linode falling behind?
Decrease exponentially? Wow, umm. No. If you drew a graph of that then people would be paying you to take their computers.
You're a funny man…
Sarcasm, or simply not proficient with exponential math?
~~http://www.vtunali.com/images/posts/hd_cost_graph.gif
~~
~~
 then the cost of computing halves. That's a negative exponential, not an exponential decrease.
Here's another way; draw a typical exponential increasing curve, y=e^x; an exponential decreasing curve (what you wrote) would be y=-(e^x). However, what you actually meant is something closer to y=e^{-x}.
@deadwalrus:
@sweh:
@deadwalrus:You yourself have acknowledged that we haven't seen legitimate hardware upgrades for about 2 years.
This was discussed before. When was the hard drive upgrade before that? 20% storage increase over 4 or 5 years is not exactly something to stand up and shout about.
Actually I documented earlier in this thread what increases there have been
Linode 64 & 1.5 GB Disk FREE 05/01/2004 - 04/30/2005 $0.00 $239.40
Linode 64 & 1.5 GB Disk FREE 05/01/2005 - 04/30/2006 $0.00 $239.40
Linode 80 & 1.5 GB Disk FREE 05/01/2006 - 04/30/2007 $0.00 $239.40
Linode 256 & 4 GB Disk FREE 05/01/2007 - 04/30/2008 $0.00 $239.40
Linode 360 & 5 GB Disk FREE 05/01/2008 - 04/30/2009 $0.00 $239.40
Linode 360 & 8 GB Disk FRE05/01/2009 - 04/30/2010 $0.00 $239.40
Linode 360 & 8 GB Disk FRE05/01/2010 - 04/30/2011 $0.00 $239.40
Linode 512 & 8 GB Disk FRE05/01/2011 - 04/30/2012 $0.00 $239.40
Linode 512 & 10 GB Disk FR05/01/2012 - 04/30/2013 $0.00 $239.40
(the free disk is 50% of the allocated disk… up until this year, where it seems the 20% addition didn't also fall into the grandfathered accounts).
Some years we get memory, some years we get disk, some years we get both. In the past 4 to 5 years my server has gone from 15Gb (5/1/2008) to 34Gb (today). So 126% increase over 4 to 5 years.
You were saying?
@sweh:
Hint: your graphs aren't drawing what you wrote. For example, your first graph is a log of a negative exponential curve, not an exponential decrease. Something totally different, entirely.
Way to try to hyper-parse to justify your math-fail. Unfortunately, your attempt is counter to the plain-and-ordinary meaning of the term "exponentially decreasing", which I submit means "decreasing according to an exponential function." Which is exactly what has happened.
Or – to put it another way -- price/gig is the inverse of gig/price, which is exponentially increasing. But whatever.
> So 126% increase over 4 to 5 years.
Glad you are happy with that. Meanwhile HDD price/gig has roughly followed Moore's law. That is, price/gig has halved every 18 months or so. Considering you appear to be challenged at calculating exponentially decreasing functions, I'll do it for you: If Linode had followed suit, you would have enjoyed a ~400% increase, give or take.
Hint: the inverse of exponentially increasing is not exponentially decreasing; that's the opposite. "Inverse" != "Opposite". You keep using words, but don't seem to know what they mean.
Ne'er mind. One day you'll learn.
And now you've gone from "20% in 5 years" to "oh, well, yeah, OK, but that's not enough".
Back in the old days of Usenet people like you were called trolls. We would have fun taunting you. I'm explicitly laughing at you because of this.
But now I'm bored. Everything you've written in response to me was according to my script. You're too predictable.
Oh how I yearn for the old days of Usenet when trolls had a couple of brain-cells…
@sednet:
Linode is being beaten on price quite badly by digitalocean and they have to respond.
…
Faster Nginx box.
25% of the price.
…
Anyway, after a bit of research I quickly discovered that Linode are in large-part running hardware that is 3-4+ years old. Which undermines the claim that DO can't sweat their assets for 5 years to generate a healthy ROI, does it not?
Secondly, I'm very lucky with Linode. The London operation seems to be rock solid and I've not had a minute's downtime. But people hosting in the US (Fremont?) don't seem to be quite as lucky. It sounds like may have had as much downtime in the last two years as the average DO customer.
@XReaper:
You just said it yourself, you pay for performance and reliability.
@sednet:Linode is being beaten on price quite badly by digitalocean and they have to respond.
…
Is Linode's performance and reliability worth 4 times more than digitalocean's? Linode has been stable as a rock but it's not like digitalocean has terrible reliability.
My DNS and mail servers have to stay up, but if my personal PBX system goes down for a few hours I'm the only one that will notice. For some things good enough at a quarter of the price is the right decision. Times are hard and money is tight, the place I work at can't even afford paper towels for the toilets.
@sednet:
the place I work at can't even afford paper towels for the toilets.
We use "toilet paper" here in the states. Using paper towels sounds like it must have been, ahhh, rough on you.
James
@zunzun:
@sednet:the place I work at can't even afford paper towels for the toilets.
We use "toilet paper" here in the states. Using paper towels sounds like it must have been, ahhh, rough on you.
James
We have toilet paper, I meant the cheap paper towels you dry your hands with after washing them.
These things:
Apparently we stopped paying the supplier so they stopped delivering.
I've had to go on raids of different floors to steal what few paper towels still exist. It's like mad max with paper towels instead of gas, staring me as Mel Gibson and nobody who can act in any of the other roles. At knocking off time we all reminisce about the before times, in the long long ago, back when we could afford paper goods.
This thread has gone impressively off-topic.
@TeshooLama:
Anyway, after a bit of research I quickly discovered that Linode are in large-part running hardware that is 3-4+ years old. Which undermines the claim that DO can't sweat their assets for 5 years to generate a healthy ROI, does it not?
Not really. From what I've noticed, Linode doesn't buy current-generation hardware. So if a server is running 3-4-year-old hardware, the server is probably 1-3 years old. That's a far cry from running for 5 years. DO also uses SSDs, which are far more expensive than Linode's hard drives, and have only recently started gaining a reputation for lasting more than a couple years.
(Hello, IRC. I loved the cat pictures, BTW.)
Edit: By the way, I don't have any particular opinion here. I just like an argument. For one thing, it's quite an assumption that DO needs 5 years to make a profit.
Edit: IRC also points out that if DO uses crappy hard drives and Linode uses ENTERPRISE ones, I'm probably wrong about the price difference.
My first month with DO has been rock solid, so I'm probably going to migrate my other services over to DO this month and cancel my Linode plan by the end of March if things still look good with everything switched over.
I will grant the DigitalOcean skeptics that if I had to place a bet on who's around in five years, I'd put it on Linode and not DO. If migrating was a massive PITA this might be relevant, but I've got my server migration procedures down to a science at this point, so if I have to switch in a year or two, it's not a really big deal, and well worth the $10-$20 a month I'll save by switching.
I don't expect the same level of support I've had with Linode over the years, but if things go poorly, I'll be happy to come back to Linode, and maybe by that time the price/performance gap will have narrowed a bit.
@Guspaz:
To use another car analogy, Digitalocean is selling a new Prius for $10k and saying they're making a profit on it.
~~
You want new Prius? I do special price. $9,995. Come with air-con, SSD, and Level 1 bandwidth.~~
@sednet:
I do special price. $9,995. Come with air-con, SSD, and Level 1 bandwidth.
You forgot to mention their corporate motto, "Me love you long time, Joe".
James
@zunzun:
You forgot to mention their corporate motto, "Me love you long time, Joe".
They used to love you long time, now they love you 1TB a month plus 2 cents a GB.
I wish Linode offered a cheap scratchpad type option for pre-production or very low intensity sites as I really wanted to stay with them. But I just cannot ignore the cost saving for a non-critical server.
Maybe they will adjust in time to offer something a bit different. I don't think large scale price cutting would be good, though a new offering with 1 core might help them keep some customers happy.
https://chesterton.id.au/state-yoda.jpg
When 800 years you are, big electric toothbrush you will need.
Who would Yoda host with?~~
@sednet:
When 800 years you are, big electric toothbrush you will need.
Yoda cleans his teeth by using the Force to teleport the bacteria into the nearby Stinky Swamp, the same way he rids himself of digestive waste. It's not like the area where he lives has a lot of paper products.
Incidentally, it was known as the Scented Swamp before Yoda showed up and teleported it full of 800 years worth of Jedi poop.
James
 apparently can die young.
@vonskippy:
~~[http://www.infoworld.com/t/solid-state-drives/test-your-ssds-or-risk-massive-data-loss-researchers-warn-213715?source=IFWNLEnltblogs2013-03-01 " target=" blank">](http://www.infoworld.com/t/solid-state- … 2013-03-01">http://www.infoworld.com/t/solid-state-drives/test-your-ssds-or-risk-massive-data-loss-researchers-warn-213715?source=IFWNLEnltblogs_2013-03-01 ](
Now that's a scary article. Thanks for the link, vonskippy.
@vonskippy:
http://www.infoworld.com/t/solid-state- … 2013-03-01">http://www.infoworld.com/t/solid-state-drives/test-your-ssds-or-risk-massive-data-loss-researchers-warn-213715?source=IFWNLEnltblogs_2013-03-01 Poor Digital Ocean - the good (and the dodgy) apparently can die young.
Wait so you are referencing an article that uses sources from 2003 of why SSDs are bad as your reason why people shouldn't use a host that use SSDs? That seems down right silly and this article is very biased. The modern smart phone (just giving a technology comparison) wasn't even released in 2003 and if they are using sources from 2003 how can this article can be trusted to compare the SSDs of today to modern infrastructure.
@chesty:
the presentation which the article is based on tested 15 currently available SSDs.
Yet they don't list any of the manufacturers. Secondly, linode/digitalocean are supposed to be in a raid 10 (which usually includes a battery backup) which should limit any data loss (which both HDDs and SSDs would have). Not only that but both should have UPS with generator power. SSDs use lower power, have a lower MTBF rate and have better performance yet Linode customers will try to find some negative to justify why they are bad? This is just silly, outright silly. This shouldn't even be considered a study.
The presentation which the article is based on listed 4 or 5 power outages suffered by major players in recent times, the UPS's failed.
I've heard SSDs fail more often than hard drives, but I haven't seen a study, feel free to cite one the shows SSDs last longer.
This is a linode forum for linode customers. You're surprised people here support linode?
The presentation and article aren't silly. The reason no manufacturers are listed are because they either want to sell that information or they can't release it due to manufactures contracts or some other reason.
DO have a slower network, provide less cores, have more VMs on a server, attract a different sort of customer that's more prone to abuse and attracts DDOS, have fewer support staff, have a smaller community and while this doesn't prove anything, they lost all of my data in less than a week. I had to rebuild from scratch.
I'm not a linode fanboi, I slam linode whenever I feel they deserve it.
@chesty:
who said DO is supposed to be raid 10? they're not. You don't know they have BBUs either, you're just guessing.
The presentation which the article is based on listed 4 or 5 power outages suffered by major players in recent times, the UPS's failed.
I've heard SSDs fail more often than hard drives, but I haven't seen a study, feel free to cite one the shows SSDs last longer.
This is a linode forum for linode customers. You're surprised people here support linode?
The presentation and article aren't silly. The reason no manufacturers are listed are because they either want to sell that information or they can't release it due to manufactures contracts or some other reason.
DO have a slower network, provide less cores, have more VMs on a server, attract a different sort of customer that's more prone to abuse and attracts DDOS, have fewer support staff, have a smaller community and while this doesn't prove anything, they lost all of my data in less than a week. I had to rebuild from scratch.
I'm not a linode fanboi, I slam linode whenever I feel they deserve it.
I'm a Linode fanboi as well and you can ask the staff how much I support them and their efforts. What I don't support is trashing other technologies and companies to make us look good.
1) You are correct neither Linode nor DO states they use battery backup in their RAID configs yet both say they use raid for redundancy. I guess we will have to trust them both since we aren't there to physically verify.
2) Wait…so…because UPS supposedly failed; SSD failed?
Lets take a quote from the article "Modern storage technology (SSDs, No-SQL databases, commoditized RAID hardware, etc.) bring new reliability challenges to the already complicated storage stack. Among other things, the behavior of these new components during power faults—which happen relatively frequently in data centers"….Wait power faults happen regularly in Data Centers? Unless you are in Fremont I would guess you would reject this statement.
3) WAT? You heard from where? Start citing some sources because SSD MTBF beats HDD MTBF outright.
4) So you question point number one saying that we don't know for sure that DO has battery backup and then make an assertion here about contracts of the test without knowing the full story? Well that seems a bit silly.
5) You are full of crap because DOs network is faster than Linodes because they don't cap their customers. I can pull many examples if you would like to prove the point (still being a huge linode fanboi).
6) Sorry to hear that. Detail your experience so the rest can be warned with facts.
2) I didn't say that, the study did though. you said DCs have UPSs, I pointed out that DCs still have power failures where the servers lose power suddenly with no chance to shut down gracefully. If you look at the presentation the article was based on, you will see a list of major players in major DCs suffering power outages recently. I can remember one, and it was amazon.
3) I heard from friends, and no thanks.
4) no you are
5) I can't believe what I'm reading, seriously, I actually lolled out loud. You see, the internet is like a series of tubes, it's not like a big truck you can just dump stuff on. The reasons you listed for DOs network being faster is the reason why it's slower, they have people on their network downloading TBs of data continuously for $5/month with NO cap, "unlimited", this slows everyone else down. Or you might be talking about linode's 50mbs out going rate? on my node in tokyo, I can burst to full 1G speed. Because linode charges a sustainable price for transfers, their network is much faster and consistently faster. They also have the money to increase capacity with demand. DO with their unlimited transfers will always be slower, their network, no matter how much capacity they add will always be saturated because people have "unlimited" transfers.
6) what details? I woke up to find my droplet was up and down and then just down, it was stuck trying to boot. I contacted support and they said they lost a raid set and my data was gone. You can't draw any conclusions from one data point though. a study from a university however…
@ohkus:
@chesty:the presentation which the article is based on tested 15 currently available SSDs.
Yet they don't list any of the manufacturers. Secondly, linode/digitalocean are supposed to be in a raid 10 (which usually includes a battery backup) which should limit any data loss (which both HDDs and SSDs would have). Not only that but both should have UPS with generator power. SSDs use lower power, have a lower MTBF rate and have better performance yet Linode customers will try to find some negative to justify why they are bad? This is just silly, outright silly. This shouldn't even be considered a study.
I also would like to know what the 2 good manufacturers are.
RAID 10 does not include a battery backup, no level of RAID specifies a battery backup. Even the battery backed write cache you get on HP servers may be useless if the SSD lies about what it has put on stable storage. UPSs, generators, and the switching equipment between them fails. The sad fact is these things are not as reliable as they are claimed to be. I've seen diesel generators fail, UPSs die after a few seconds of runtime, switching equipment get stuck on or off. There is a reason EMC puts batteries in each and every symmetrix array.
SSDs fail in normal operation and when they do you don't even have the option of recovering anything, it's just all gone in one hit with no advance notice. DO better be using raid1 but I suspect they don't due to the extra costs involved.
How many times has a disk failed in a Linode host, Linode replaced it, and no customers even knew about it? I bet it happens all the time.
@chesty:
I've heard second hand DO use raid 5
…
I contacted support and they said they lost a raid set and my data was gone.
I do hope they are not cutting costs by using raid 5 on consumer grade SSDs.
@ohkus:
Wait so you are referencing an article that uses sources from 2003
Huh?
The paper was written for FAST '13 (the File and Storage Technologies Conference held in FEB 2013).
From what ass did you pull the 2003 figure?
The paper was done by postdoc's at OSU (Ohio State University) and by researchers at HP Labs.
They used 15 MODERN SSD drives from 5 MODERN vendors.
Geesh, the Infoworld article linked to all the technical papers - how hard was it to click on a few links and get your facts straight.
@vonskippy:
@ohkus:Wait so you are referencing an article that uses sources from 2003
Huh?The paper was written for FAST '13 (the File and Storage Technologies Conference held in FEB 2013).
From what ass did you pull the 2003 figure?
The paper was done by postdoc's at OSU (Ohio State University) and by researchers at HP Labs.
They used 15 MODERN SSD drives from 5 MODERN vendors.
Geesh, the Infoworld article linked to all the technical papers - how hard was it to click on a few links and get your facts straight.
However, while we have over 50 years of collected wisdom working with spinning disk, flash-based SSDs are relatively new [1]
[1] Roberto Bez, Emilio Camerlenghi, Alberto Mod- elli, and Angelo Visconti. Introduction to Flash Memory. In Procedings of the IEEE, pages 489– 502, April 2003.
Go look at the sources the article uses….
If you don't understand how science (and engineering) papers are written, stick to the Unoobtu forums.
What part of "Introduction" is unclear - they site a reference so they don't have duplicate explaining to motards the basics of how flash memory works. The SSD tests were done in 2012. If you're going to spread FUD, at least do it right.
OMG - the paper uses DATA from 1990 - yes 1990 - yes 33 freaking years ago - so of course the study must be completely dated and bogus (the 1990 paper was by Gibson explaining the limitations of RAID, and it was his PhD thesis at UC Berkeley - Go Bears!).
Sources: State of Iowa Department of Health, Record of Live Birth, 1980. Wegmans Bakery, Happy 0x20th Birthday Hoo Pycat, 2012.
@jed:
I skipped pages 4 through 11 of this thread. Did I miss anything?
No.
@jed:
I skipped pages 4 through 11 of this thread. Did I miss anything?
Page 11 has a discussion of Yoda's bowel movements, surely that should be reviewed to ensure Linode's servers are running at optimim levels of performance.
James
@ohkus:
@vonskippy:
@ohkus:Wait so you are referencing an article that uses sources from 2003
Huh?The paper was written for FAST '13 (the File and Storage Technologies Conference held in FEB 2013).
From what ass did you pull the 2003 figure?
The paper was done by postdoc's at OSU (Ohio State University) and by researchers at HP Labs.
They used 15 MODERN SSD drives from 5 MODERN vendors.
Geesh, the Infoworld article linked to all the technical papers - how hard was it to click on a few links and get your facts straight.
However, while we have over 50 years of collected wisdom working with spinning disk, flash-based SSDs are relatively new [1]
[1] Roberto Bez, Emilio Camerlenghi, Alberto Mod- elli, and Angelo Visconti. Introduction to Flash Memory. In Procedings of the IEEE, pages 489– 502, April 2003.
Go look at the sources the article uses….
You're showing your ignorance of how a technical paper is written, not poking holes in the paper itself. They're referencing a 10 year old paper to show that "flash-based SSD's are relatively new".
@bryantrv:
I keep all my important data on 8" floppies 'cause that's what Jerry said to do in Byte.
My work still has some data on Travan 4 tapes. At least we think it's on there, we don't have a working drive to read them.
 13 of 15 drives suffered failures after cutting the power 3000 times. Not a single time, but 3000 times. When is that last time your Linode suffered an unexpected power failure-based shutdown? It seems to only happen every few years. 3000 is an insanely high figure even for consumer use, let alone datacenter use.
2) Of the problems reported, only one failed entirely, others merely suffered some sort of problem, such as data corruption. How much of this was because of the SSD, and how much was because of the filesystem/OS?
3) They don't appear to have performed a similar test on magnetic HDDs as a comparison. For all we know, if you took 15 HDDs and subjected them to the same tests, you'd see similar results.
EDIT: Actually, reading the paper, they do some limited testing on HDDs (a single 5.4K RPM and a single 15K RPM model). One of the HDDs (the 5.4K RPM one, I believe it was) didn't suffer failure (similar to one of the SSDs), and the 15K RPM one did suffer failures. Still, none of their test results differentiate between physical errors and filesystem errors. It seems like the vast majority of errors suffered would be solved by a copy-on-write filesystem with per-block checksums and some sort of redundancy (even if it's a single disk with ditto blocks).
It's always good to have alternatives. I am glad DO came about, gives Linode more reason to be competitive with their offerings.
Not so good for DA
@Nuvini:
Just saw this:
http://www.wired.com/wiredenterprise/20 … italocean/">http://www.wired.com/wiredenterprise/2013/04/digitalocean/
Linodes used to have the same problem - if you created a raw disk image it would be full of other people's data. I haven't tested it lately, but they say it isn't so anymore. I'm still paranoid enough to zero out partitions myself before deleting them.
@Nuvini:
Just saw this:
http://www.wired.com/wiredenterprise/20 … italocean/">http://www.wired.com/wiredenterprise/2013/04/digitalocean/ Not so good for DA
:x
Someone should tell this Kenneth White about cat abuse
cat /dev/vda | strings > /dev/shm/dump.txt
Totally and utterly pointless use of cat. It should be:
strings /dev/vda >/dev/shm/dump.txt
And maybe don't store massive files in a tmpfs.
> “Formatting the drive before I get it; yeah that would be good.”
And he doesn't seem to know the difference between formating and blanking either.
Now that it's online - it's complete drivel, written by fops that are clueless (at best).
Wired is now up there with "the Sun" and "the National Enquirer", just targeted at hipsters instead of the blue hair crowd.
OK, so I've been procrastinating that for a few years now.
@Nuvini:
Just saw this:
http://www.wired.com/wiredenterprise/20 … italocean/">http://www.wired.com/wiredenterprise/2013/04/digitalocean/ Not so good for DA
:x
For what it's worth I created a 4GB droplet in Amsterdam and ran strings on the disk. The only password hashes or private keys I could find came from my newly created filesystem. It's looks like they really have fixed the problem.
I'll try the same on an EBS volume from amazon just for fun.
EDIT: Amazon EBS volumes contain nothing but zeros.
EDIT2: Amazon Instance storage contains nothing but a filesystem over zeroed blocks.
Am a desktop gnu/nix fanatic, the logical next step seemed to be networking and server admin and so here I am. Atm have neither an active acct with either provider .. Linode or Digitalocean. Still window shopping right now. All that said … opinions follow.
Wouldn't go so far as to say Digitalocean is a Linode "competitor", for many reasons.
a. They don't have the track record, thusly the proven reputation Linode has acquired and kept up.
b. They have a presence in all of 2 datacenters last time I checked, 1 in NYC ( US ) and one in Europe …( Linode has 6, here in the States.)
c. Already DO apparently has had more than a few "technical issues", like losing clients ( at the time … free backups.) Not allowing virtio by default, apparently it has to be specifically requested by clients.
d. Imo … saying provider has ssdrives, leaves A LOT of speculation and/or territory to cover. Ssd's no doubt come in many varied qualities. While it may seem off the top attractive, still leaves many questions in my mind. Ssd's imply faster disk i/o … But is that so on a box that may be highly oversold ? Guessing any good server admin, will try to strictly limit disk i/o anyway, as a standard operating procedure. Also ssd's are not esp known for graceful failure. Buncha other stuff comes to mind on this pt alone.
Conclusion: Based solely on mucho research and/or common sense. Personally wouldn't go so far as to classify digiocean as a Linode competitor. In my mind, at the time, DO is more a cheap source of vps's I'd feck around with to learn and once the skillset is up to par, move on into more reputable providers like Linode for real world production level stuff.
O course … could always be wrong/etc. Tend to try to readily admit this simple statement of fact.
Main reason, for including the "could be mistaken" disclaimer. Though still seems Linodes got 4 to 1 options in US. Which is the target audience atm I have any reason to care about, shrugs. For real … am not advocating either/or … That's x-persons choice to make.
Only personally doubting, that someone like digitalocean should be considered/count as a full fledged "competitor" for people like linode.
Yet again … admit, I could be wrong !
@anyusernamehere:
Disclaimer 1st …
:) b. They have a presence in all of 2 datacenters last time I checked, 1 in NYC ( US ) and one in Europe …( Linode has 6, here in the States.)
A bit of a nit; I know their promotional material states New York, but I believe it is in fact North Bergen, NJ. Admittedly, there's not a lot of difference than New York, NY but neither is between Newark, NJ and New York, NY. I just think providers should be upfront and clear about their locations and where customer data is physically stored.
@AVonGauss:
@anyusernamehere:Disclaimer 1st …
:) b. They have a presence in all of 2 datacenters last time I checked, 1 in NYC ( US ) and one in Europe …( Linode has 6, here in the States.)A bit of a nit; I know their promotional material states New York, but I believe it is in fact North Bergen, NJ. Admittedly, there's not a lot of difference than New York, NY but neither is between Newark, NJ and New York, NY. I just think providers should be upfront and clear about their locations and where customer data is physically stored.
You're correct about NJ. Not sure about the city, but their marketing about New York is a lie. (or was).
@AVonGauss:
A bit of a nit; I know their promotional material states New York, but I believe it is in fact North Bergen, NJ. Admittedly, there's not a lot of difference than New York, NY but neither is between Newark, NJ and New York, NY. I just think providers should be upfront and clear about their locations and where customer data is physically stored.
Meh. Linode's "Newark" location is actually in Cedar Knolls. "Dallas" is in Grapevine! Who cares? At the speed of light it's not a bit deal (for most purposes, anyway).
@mnordhoff:
[Meh. Linode's "Newark" location is actually in Cedar Knolls. "Dallas" is in Grapevine! Who cares? At the speed of light it's not a bit deal (for most purposes, anyway).
[/quote] I stated as much in my first post, and I agree regarding Linode's datacenters, but at least they are in the same state advertised. There's not much difference in the latency between the locations, but which jurisdiction it falls under can and is increasingly so important to know for businesses. Besides, what else "sounded better" than the actual truth?
@AVonGauss:
I stated as much in my first post, …
Oh. So you did.:oops:
@AVonGauss:
… and I agree regarding Linode's datacenters, .but at least they are in the same state advertised. There's not much difference in the latency between the locations, but which jurisdiction it falls under can and is increasingly so important to know for businesses. Besides, what else "sounded better" than the actual truth?
I'm rather fond of organizing by IATA code, which puts Linode's Newark location under Newark (I think) or NYC. DigitalOcean's location would probably be similar.
I'd honestly never thought of state boundaries mattering for business/legal reasons, but that makes sense. IATA codes are so fuzzy (and proprietary!) anyway.
/me shrugs.
I once had a client with their offices in Amarillo, TX require that the server must not be in California or New York, had to be within the US, but had to be at least 500 miles from their offices so Dallas was not an option. The only sites that fit the requirement were the Atlanta, GA DC or the "New York" DC. Had to go with Atlanta since even though the NY site was actually in NJ, since the "Suits" had a fit since the documentation had "New York" on it.
This was for a "hot site" [no not that kind…
@peleus:
Does some here have a review with DigitalOcean? I would like to ask please on how their tech support is particularly the response time.
Terrible compared to Linode
@vonskippy:
Q: Do you limit bandwidth? A: No we do not limit monthly bandwidth in anyway. We allow you to fully utilize the network and burst traffic without limits.
Yeah, that always works out well because nobody would ever think of abusing that and hogging all the bandwidth for themselves.
That right there is a reason not to go with them. Dreamhost VPS is unreliable for the same problem. All that "unlimited" means is that they aren't going to tell how how much bandwidth you get. Nothing is unlimited. The way those promises work is their average bandwidth use per customer is something like 1TB (totally hypothetical), they have no problem keeping up with this level of usage, and so they say that they offer "unlimited" bandwidth because they know most customers won't exceed what they can supply. If sites use more, the hosting company might just ignore them. They could also even cancel service to websites that use too much bandwidth like shared hosting providers, which negates one of the major benefits of VPS.
@kbar:
imo just because they're sponsored by a vc doesn't mean they're legit.
in fact, I would be more worried than in 6 months or so they would get sold out to some other company.
i would also be weary as they're hitting all the keywords (cloud, ssd, unlimited bw, pay-as-you-grow, etc)
Same reasons as mentioned above for the buzzword use. You want to stay away rom anything that tries to sound too awesome; it seems very similar to GoDaddy. Company transfer and oblivion is a serious issue. I did a lot of work for Wikkii only to have the service transferred to a company called "Gunggo," which has apperently ditched the service.
The only real problem I had with Digital Ocean was that their mother was a hampster and their father smells of elderberries.
I give them 4 out of 5. The main reason being it is kind of annoying that you can't get larger disk space for a better value of money.
@depill:
I give them 4 out of 5. The main reason being it is kind of annoying that you can't get larger disk space for a better value of money.
To be fair, Linode has this exact same problem. I'm hopeful that the advent of the hybrid SSD linodes will allow Linode to switch from 15K RPM drives (which max out at 600GB) to 7200RPM drives (which max out at 4000GB). The SSD caching should alleviate any performance concerns while giving us more usable space.
I'm hosting a bunch of footage from our convention on my personal linode because our company one doesn't have the space, and it'd be nice to be able to host footage from previous events too…
@autone:
I still have my DO droplet till today and it's great value for money. Linode and DO have their own clients and fill different niches.
Totally agree. I am back with Linode for my main stuff, but DO can't be beaten (imho) for the lower end. Beefier stuff I will stick with Linode, but it's nice to have a per-hour based scratchpad to try stuff out on.
@sblantipodi:
it seems really interesting
Why? You can install VirtualBox, VMware or use KVM for free and customize and run 24 hours a day for free.
The disks performance are incredible, I can easily read up to 250MB/sec while I can't go over 100MB/sec with linode.
I like the possibility to produce a manual snapshot without the needs of paying a backup service.
I'm really disappointed from the fact that I cannot use my own kernel on digitalocean since they doesn't support something like pv-grub.
I like the possibility to create a linux image in my virtual box and load it on a linode, I cannot do this on DO.
My linode uses a Xeon E5 2670, don't know what DO uses.
I "feel more freedom" with linode and I like linode more, it costs more but it offer more.
@Guspaz:
Linode's new hybrid SSD instances that are in beta testing should resolve the disk performance gap.
nice to know that linode is working in this sense. is there any ETA for the switch from HDD to SSD?
@Guspaz:
Caker said they'll make an announcement on the subject in Q1 of 2014.
thanks for the answer, I hope that the "storage capacity" will be the same.
@sblantipodi:
@Guspaz:Caker said they'll make an announcement on the subject in Q1 of 2014.
thanks for the answer, I hope that the "storage capacity" will be the same.
Check in the beta forum, it sounds like it'll be SSD's acting as a cache for HDD. So, same storage from the HDD backend, but increased speed from the SSD cache.
@glg:
@sblantipodi:
@Guspaz:Caker said they'll make an announcement on the subject in Q1 of 2014.
thanks for the answer, I hope that the "storage capacity" will be the same.
Check in the beta forum, it sounds like it'll be SSD's acting as a cache for HDD. So, same storage from the HDD backend, but increased speed from the SSD cache.
ah ok, an hybrid solution.
I pay twice for linode than digitalocean, can't understand why digital ocean offers SSDs (not hybrid solution but SSDs) while linode with double the price offers a simple hybrid solution.
@sblantipodi:
I pay twice for linode than digitalocean, can't understand why digital ocean offers SSDs (not hybrid solution but SSDs) while linode with double the price offers a simple hybrid solution.
My wife says the same thing when comparing my Audi to her Honda.
There's cheap and then there's good, you seldom (if ever) get them both at the same time.
@caker:
You'll have to just wait and see.
-Chris
we'll wait.
@sblantipodi:
ah ok, an hybrid solution.
I pay twice for linode than digitalocean, can't understand why digital ocean offers SSDs (not hybrid solution but SSDs) while linode with double the price offers a simple hybrid solution.
Your moment of Zen will come when you realize that of the tens of thousands of files on a VPS, your working set of data is all that you are really touching. Sure, you occasionally fiddle with a log or a binary, but access pattern for a Web server for example will almost entirely revolve around the data being served.
From there, you extrapolate "do I really need /var/lib/games on a superfast SSD?" and you shall become enlightened. And it is good.
@jed:
@sblantipodi:ah ok, an hybrid solution.
I pay twice for linode than digitalocean, can't understand why digital ocean offers SSDs (not hybrid solution but SSDs) while linode with double the price offers a simple hybrid solution.
Your moment of Zen will come when you realize that of the tens of thousands of files on a VPS, your working set of data is all that you are really touching. Sure, you occasionally fiddle with a log or a binary, but access pattern for a Web server for example will almost entirely revolve around the data being served.From there, you extrapolate "do I really need /var/lib/games on a superfast SSD?" and you shall become enlightened. And it is good.
If I need it or not is not a problem, I pay twice than digital ocean, give me some reason
@sblantipodi:
If I need it or not is not a problem, I pay twice than digital ocean, give me some reason
:)
Then move - you do nothing but whine - if you're so very unhappy with the service Linode provides at the price they charge, leave. It's not a complex problem.
@vonskippy:
@sblantipodi:If I need it or not is not a problem, I pay twice than digital ocean, give me some reason
:)
Then move - you do nothing but whine - if you're so very unhappy with the service Linode provides at the price they charge, leave. It's not a complex problem.
you never seen a person whining.
I like linode this is the reason why I'm here. if you don't like to read my posts just skip them.
@Benio:
****Post subject:
New Linode competitor - DigitaloceanPosted:**** Tuesday, January 15, 2013, 18:40:30
This calls for…! DigitalCake
~~
Or possibly…an AlcoholOcean!
![](~~