The New Linode Cloud: SSDs, Double RAM and much more

https://blog.linode.com/2014/04/17/linode-cloud-ssds-double-ram-much-more/

by qmr • 12 years ago

584 275 12 years ago

blog.linode.com

giulianob5• 12 years ago

Holy crap this is awesome. Good job guys at Linode. I said I would switch if the prices dropped about 25% because RAM was pricey.... So now I have to switch.

extesy• 12 years ago

So now they match DigitalOcean prices but offer slightly more SSD space for each plan. I wonder what DO answer to this would be. They haven't changed their pricing for quite a while.

catinsocks• 12 years ago

DigitalOcean still beats Linode on price as half the time I don't need a 2GB server but I do want to keep servers separate (multiple 512s at $5 is better for me than 1 2GB).

giulianob5• 12 years ago

I need 2gb - 4gb servers and DO's quality has decreased since they've grown so much.

Jake2322• 12 years ago

The increased CPU on some of the plans compared to DO is what is encouraging me to change. When doing web scraping, it's easy to max the CPU once you are doing 20 paths on a page.

rjknight• 12 years ago

It looks like Linode are still leaving the "incredibly cheap tiny box" market to DO. Linode's cheapest option is $20/month, which makes it slightly less useful for the kind of "so cheap you don't even think about it" boxes that DO provide.

EC1• 12 years ago

Shit, I should probably move to DO. I've been running a Linode for like two years, and have only ever used it once, yet I keep getting billed.

eli3• 12 years ago

Well, if you're truly not using it, you could delete the instance and not be billed, then redploy when you need it.

trumbitta2• 12 years ago

They introduced hourly billing like a week ago :)

kylec3• 12 years ago

Yeah, my current Linode I've been running since it was a 512MB instance. I definitely don't need 2GB, I'd rather pay a bit less per month than get the upgrade. I don't think they need to go down to $5/month, but a $10/month plan similar to DO's would be very useful.

morganherlocker2• 12 years ago

I use a tiny VPS provider that is actually $20/yr. For sites that get a couple thousand hits per month it is perfect. Although I do not see it as a requirement in this price range, the uptime has been on par with the big names (I have known of one 5 hour outage in the 3 years I have been using it). Even DO's $5/mo now seems a bit steep now for the "I just had a random idea" type thing.

kbar13• 12 years ago

5c a month savings is worth keeping your Linode on old hardware with 512MB of ram?!??!?

TillE• 12 years ago

the "incredibly cheap tiny box" market

How big is that market, really? I keep a small VPS around for whatever (I like Hetzner), and I'm sure lots of nerds are the same. But for any usage that's even a little bit serious, $20/month is peanuts.

bluedino3• 12 years ago

Probably pretty big - look at how popular tiny VPS providers are.

A lot people don't need an entire server and just need a shell, some file space, and a small web/app server.

Even just 10 years ago it was very expensive to get a non-shared hosting account. Go back a little further than that, and you were elite if you had a real server out there somewhere. Back then a low-spec co-located server was $300 a month. And let's not even talk about how expensive bandwidth was!

kylec3• 12 years ago

What if you want to have several tiny boxes? For example, if you want to host a mail server as a separate box for security, that's another $20/month. Or what if you need to use a different Linux distro for something? Another $20/month. A few examples like that add up fast.

lelandriordan• 12 years ago

I bet its a lot bigger than you think. For example, you can handle a ton of views on something like WordPress or Ghost with DO's smallest offering. Especially if you use it in combination with caching and and a reverse proxy like Cloudflare.

mcone2• 12 years ago

One reason Linode doesn't offer "tiny boxes" is that they, like other companies, are worried about running out of IPv4 addresses.

tszming• 12 years ago

This is not true.

Linode allow you to attach additional IPs to your Linode at $1/month, for example, I have one Linode which has 6 IPs attached for only $25/month. I still can't find a company has this kind of flexibility.

thisishugo• 12 years ago

I doubt that's true. More likely it's because lower tier users are almost always disproportionally more demanding of help than higher tier ones. If you're going to make, for example, $1,000/m you're better off getting it from two competent, $500/m customers that don't need much help, rather than 100 $10/m customers who all need their hands held constantly. In Linode's case fewer users also makes contention less of an issue - the fewer VPSs per physical box, the better.

bluedino3• 12 years ago

$2 boxes are ripe for abuse and billing fraud.

voidlogic2• 12 years ago

Linode's cheapest option is $20/month, which makes it slightly less useful for the kind of "so cheap you don't even think about it" boxes that DO provide.

Yes, this, the $10 plan is DO's most popular.

dotBen• 12 years ago

How do you know DO's $10/m plan is most popular?

(Please don't say it's because the pricing page highlights the $10 plan ass "most popular". You know that's just a classic up-sell trick, right?)

meritt• 12 years ago

I have a feeling it's a bit of customer service / risk mitigation. Obviously there are numerous exceptions but the sort of people that execute a lot of fraud, DDOS'ing, or simply amateurs playing around are also the sort of people that'll use DO before Linode. I imagine there's a heavy cost associated with hosting those users relative to ones happy to pay $20/mo instead.

brandon2722• 12 years ago

As someone who has used both Linode and "low end box" services, both have their place. Two observations:

1) I feel like Linode offers a level of support that the others don't offer. There are lower end providers that do offer good support, but it's not as consistently good as Linode. Were they to start offering cheaper $5 or $10 a month plans, they would be challenged when it comes to continuing to offer that level of support.

2) Observationally, it seems to me that Linode customers on average tend to be more serious about their usage than those who are looking for a <$5/month VPS. Go to the IRC channels for each company and see who hangs out there. That said, this should only matter to you if you are concerned about who your neighbours may be on a VM host, if you plan on being part of the IRC community, or if you are concerned about how reputable your IP address looks given what block it belongs to.

filmgirlcw3• 12 years ago

I agree with this. I have a few DO accounts, a Linode and a few BuyVM accounts. The Linode is what I use for all of my "serious" production stuff, but I love the LEBs to play.

That said, I do think the performance of DO is quite good. It reminds me of Slicehost back in the day.

The BuyVM guys are just the best people ever, and I mostly maintain play boxes there just b/c I can get away with stuff there that might not fly at DO or Linode.

filmgirlcw3• 12 years ago

And I would imagine that is very much by design. DO's model is to scale as much as possible -- they want to be all over the place and they have the funding to make those sorts of changes.

As a more established provider, Linode has to decide if adding more customers would be worth the incremental support costs. Given Linode's reputation for support, adding a bunch of LEB customers who aren't likely to renew at the same rates as people who pay more money - and are also likely to have more support requests - competing with that market doesn't make a ton of sense.

Plus, whether fair or not, having a minimum pricing point makes a service seem more premium -- more "ready for serious business." Now, obviously we've all been with expensive hosts that have terrible performance -- but you'd be surprised how many non-technical decision makers will choose a higher-priced offering over a lower-priced offering, on the supposition that "more expensive = better."

the_ancient• 12 years ago

The type of people that will be buying a $5 server are the type of people that will be sending in 10000000 help tickets on how to do X on that server

davidgerard• 12 years ago

There's a bit of getting what you pay for. That said, Rackspace are in the same sort of market tier as Linode, and in my experience of both Linode are just ridiculously better. I would heartily recommend Linode to anyone.

relaxatorium• 12 years ago

This seems pretty fantastic, I am excited to upgrade and think the SSD storage is going to be really helpful for improving the performance of my applications hosted there.

That said, I am not an expert on CPU virtualization but I did notice that the new plans are differently phrased than the old ones here. The old plans all talked about 8 CPU cores with various 1x, 2x priority levels (https://blog.linode.com/2013/04/09/linode-nextgen-ram-upgrad... for examples), while the new plans all talk about 1, 2, etc. core counts.

Could anyone with more expertise here tell me whether this is a sneaky reduction in CPU power for the lower tiered plans, or just a simpler way of saying the same thing as the old plans?

Jake2322• 12 years ago

Somebody asked this on the blog post, here was the reply from Linode:

If you take the upgrade, you inherit the new plan specs, vcpus and all.

We’ve greatly reduced the contention on these new machines compared to our old structure, and in testing this new arrangement provides much more consistent CPU time with less potential for steal. We think it’s great and totally worth the move, otherwise we wouldn’t have done it. These machines are incredibly fast, faster procs, SSDs, the network is incredible, etc.

jaequery5• 12 years ago

cpu counts dont mean much in terms of performance. but from a strict performance standpoint, their E5 2680's are amongst the tops in the industry (http://www.cpubenchmark.net/high_end_cpus.html). i'm pretty sure DO's is somewhere around 1/10th the performance of Linode's in terms of CPU power. you are getting a really good deal here w/ Linode, I dont think any cloud providers are at this level right now.

ac29• 12 years ago

Amazon also uses E5-2680 v2 on their C3 instance types, and E5-2670 v2 on R3 types.

izietto• 12 years ago

Do you know cheaper alternatives? Like DigitalOcean, as @catinsocks suggests

nivla2• 12 years ago

Awesome News. Competition really pushes companies to please their customers. Ever since Digital Ocean became the new hip, Linode has been pushing harder. My experience with them has been mixed. Forgiving their previous mishaps and the feeling that the level of Customer Service has gone down, they have been decent year long. I wouldn't mind recommending them.

[Edit: Removed the bit about DigitalOcean Plans. If you have Ghostery running, it apparently takes out the html block listing different plans]

mimiflynn• 12 years ago

https://www.digitalocean.com/pricing/

cshenoy• 12 years ago

You mean https://www.digitalocean.com/pricing/ ? Seems to be working at the moment.

nivla2• 12 years ago

I am really sorry. I had Ghostery running, it took out the entire plan block so I was only able to see the FAQ.

svmegatron• 12 years ago

It seems to be the Optimizely js, which could mean they are A/B testing something about their pricing grid.

I'd really like to know the results of the test.

codecondo• 12 years ago

What tests? I doubt it's any of your business.

jaequery5• 12 years ago

im really impressed by their new CPU specs. from experience those aren't cheap and it's possibly the fastest CPU out in the market. combined with the SSDs, it may be that Linode currently is the fastest of any cloud hosting right now.

h4pless2• 12 years ago

I notice that Linode talked a good bit about their bandwidth and included outbound bandwidth in their pricing model which DO does not. I wonder if DO has a similar model or if transfer capacity the only thing you have control over.

geekam• 12 years ago

I do not understand the difference between outbound bandwidth vs the transfer. Can you explain a bit please?

honopu• 12 years ago

Bandwidth - How big the pipe is.

Transfer - How much they let through the pipe.

Just a guess.

kilburn• 12 years ago

If you read below the price tables you'll get a good hint:

What happens if I exceed my monthly transfer quota?
You will be invoiced $0.10 for each GB over your pooled network transfer quota. Please note that all inbound traffic is free and will not count against your quota.

Bandwidth is the maximum Mbps that your VPS can spill out at any moment, whereas transfer is the maximum amount of per-month cumulative traffic that you can send out before incurring on additional traffic costs.

h4pless2• 12 years ago

Yeah, honopu and kilburn got it right. My understanding is that transfer measurements are the maximum amount of data your instance can serve up in a month, while bandwidth is the speed at which it can do it.

Its like downloading a 1 GB file to your computer using a 56k modem or a cable modem... the transfer amount will be 1 GB either way, but the cable modem will be a hell of a lot faster due to the increase in its bandwidth. As you upgrade your plan with Linode, your "modem" gets faster, however DO doesn't mention the speed of their "modem" or if its speed is affected by the price level you select.

JohnnyBrown• 12 years ago

bandwidth is the derivative of transfer over time. It has the same relationship to transfer as speed has to distance.

bandwidth = d(transfer) / dt

gnopgnip• 12 years ago

Bandwidth is cheap when you are in a datacenter.

endijs2• 12 years ago

Most interesting part in this great upgrade is that they went from 8CPU setup to 2CPU setup. But yeah - 2x more RAM, SSDs will guarantee that I'm not going to switch anytime soon. Sadly I need to wait a week until this will be available in London.

eli3• 12 years ago

"vCPU"

In a virtualized environment, the number of "CPUs" you can see is probably not a great way to measure how much computing power you have.

stefantalpalaru• 12 years ago

On a host that's mostly idle and where there are no CPU usage caps, seeing 2 cores instead of 8 is a pretty serious drawback if you use Gentoo on Linode like I do.

andypants2• 12 years ago

They did upgrade the cpus though - "Linodes will now receive Intel’s latest high-end Ivy Bridge E5-2680.v2 full-power server-grade processors."

endijs2• 12 years ago

Yes. That's good. And Caker also wrote in blog comments: "We’ve greatly reduced the contention on these new machines compared to our old structure, and in testing this new arrangement provides much more consistent CPU time with less potential for steal. We think it’s great and totally worth the move, otherwise we wouldn’t have done it." So by their estimates 2 now should outperform previous 8.

spyder• 12 years ago

One comment on their blog post says that after the change he is getting lower performance on a CPU-heavy test (vanitygen):

https://blog.linode.com/2014/04/17/linode-cloud-ssds-double-...

IO is probably more important for most of the peoples who use VPS, but It would be good to see CPU test from others too, not just IO benchmarks.

bfrog• 12 years ago

I'm actually a little unhappy, it looks like they reduced the CPU count for my $20/mo instance. At this point there's basically no reason to stay with them now.

eli3• 12 years ago

I would encourage you to run a benchmark before deciding that 2 vCPUs is necessarily worse than 8 vCPUs running in a different architecture.

ckozlowski3• 12 years ago

In a VMware environment, I'm often educating my users that more vCPUs doesn't necessarily equate to higher performance, as the delays incurred in co-scheduling increase with the number of vCPUs.

For VMs that are not hypervisor-aware, the hypervisor must have all of that VM's vCPUs available at execution time since it cannot always anticipate which vCPUs the guest OS will execute against.

With fewer vCPUs, one might see what you're alluding to; better utilization and lower wait states because it's easier to schedule 2 CPUs consistantly than 8 CPUs on a host running multiple VMs.

Of course, that's in a VMware environment, so forgive me if Linode's hypervisor of choice avoids this problem I'm sure they're using KVM or Xen or something along those lines, maybe they're not as susceptible.

opendais3• 12 years ago

tl;dr: If you had 40 VMs on 8 cores before, you have 10 on 2 cores now. It is the same ratio of VMs:Cores but with stronger processors.

Long version: "If you take the upgrade, you inherit the new plan specs, vcpus and all.

From Caker's comment on the blog. It seems that this was done to reduce fighting over core and provided more consistent fair availability of processing power when they tested it b/t VMs.

ckozlowski3• 12 years ago

Makes a lot of sense.

If your hosts have, say, dual 4-way CPUs, and you're giving your VMs 8 vCPUs, then a single VM can execute per clock cycle since a VM needs all of it's vCPUs made available to the guest OS. With 8 VMs, that means one VM is executing every 8th clock cycle.

If you "downgrade" those VMs to 2 vCPUs, then 4 VMs can execute per clock cycle, and that VM can now execute every other clock cycle instead of waiting 8 cycles. More work gets done, even though the amount of CPUs has gone down.

Since most VMs are probably executing work that only has 1-2 threads, then there's no loss from lack of parallelism. Remember, a VM needs all cores available, regardless of how many threads need to be executed.

Of course, some workloads do need that many threads, and so balancing number of cores vs. available execution time becomes a little more tricky. but based on what you shared, it sounds like Linode took a closer look and came to the same conclusion.

SudoAlex2• 12 years ago

This might be a good thing for some customers.

In the past on certain host servers you get a great experience - low latency and the ability to burst CPU usage. However occasionally you get provisioned on an awful server where the latency of your server varies significantly, and your apps struggle to get a decent share of the CPU.

I'd rather see my servers have a small but more fair amount of burst instead of all packages getting 8 CPUs with some of them getting a bad experience from noisy neighbours.

davexunit• 12 years ago

Cool news, but their website now has the same lame design as DigitalOcean. I liked the old site layout better.

kawsper3• 12 years ago

Me too, but fortunately their administration system is still old-school.

I think their platform needs some love, it is a bit clunky to work with sometimes. But they are my favorite host.

6cxs2hd6• 12 years ago

Looks like library.linode.com has the same format, and is still the same fantastic resource.

I'm embarrassed to say that once upon a time I used library.linode.com to help walk me through my first Linux server setup... running on Digital Ocean. Yes, I felt like a complete douchebag. I wanted to use Linode but I was one of those $5/month people.

However it left me with a sense of obligation to give Linode business someday when I could justify/afford doing so. It looks like Linode just gave me the opportunity to do that.

kevinqi• 12 years ago

Yeah, the library is really excellent - easy to follow and very extensive. It's smart of them to do paid crowdsourcing to create the articles.

It's cool of Linode to make the library available to everyone, not just customers.

thejosh• 12 years ago

DO do the same thing, plus they get paid for articles.

Shame browsing it since the redesign is a complete clusterfuck between articles and community forum posts!

morninj• 12 years ago

What makes it lame?

ilaksh• 12 years ago

Yeah, I don't think its lame at all. Its much more modern and nicer.

But that is funny, they copied so many aspects of Digital Ocean's design.

kh_hk2• 12 years ago

- 2003: http://web.archive.org/web/20030614084310/http://www.linode....

- 2007: http://web.archive.org/web/20071214070928/http://www.linode....

- 2013: http://web.archive.org/web/20131228003146/https://www.linode...

- 2014: https://www.linode.com/

Give me my beveled borders back!

Justen• 12 years ago

Higher specs sound really nice, but on HN I see people commenting on the ease of DO's admin tools. How does Linode's compare?

kawsper3• 12 years ago

The interface looks a little more clunky, but it works, and it is easy to work with.

noir_lord• 12 years ago

It's not clunky it's more "dated" (and very functional).

It also does a fair bit more so apples to oranges a little bit.

mwexler• 12 years ago

There's similar and then there's alike. I guess it makes comparison easy, but imitation certainly must be the sincerest form of flattery:

Compare the look and feel of https://www.linode.com/pricing/ and https://www.digitalocean.com/pricing/

jasonlotito• 12 years ago

And DO's looks like a million other plan pages that already exist. It's not some unique design to DO.

jevinskie• 12 years ago

I resized a 1024 instance to 2048 last night and it looks like it is already running on the new processors (from /proc/cpuinfo): model name : Intel(R) Xeon(R) CPU E5-2680 v2 @ 2.80GHz

Should I upgrade? Do I want 2 x RAM for 1/2 vCPUs? =)

raverbashing• 12 years ago

Congratulation on Linode

I stopped being a customer since migrating to DO but my needs were really small

But I think their strategy of keeping the price and increasing capabilities are good. Between $5 and $20 is a "big" difference for one person (still, it's a day's lunch), for a company it's nothing.

However, I would definitely go to Linode for CPU/IO intensive tasks. Amazon sucks at these (more benchmarks between the providers are of course welcome)

kyrra3• 12 years ago

I forgot to benchmark the disk before I upgraded but here are some simple disk benchmarks on an upgraded linode (the $20 plan, now with SSD)

  $ dd bs=1M count=1024 if=/dev/zero of=test conv=fdatasync
  1024+0 records in
  1024+0 records out
  1073741824 bytes (1.1 GB) copied, 1.31593 s, 816 MB/s

  $ hdparm -tT /dev/xvda
  /dev/xvda:
   Timing cached reads:   19872 MB in  1.98 seconds = 10020.63 MB/sec
   Timing buffered disk reads: 2558 MB in  3.00 seconds = 852.57 MB/sec

Upgraded cpuinfo model: Intel(R) Xeon(R) CPU E5-2680 v2 @ 2.80GHz

Old cpuinfo model: Intel(R) Xeon(R) CPU L5520 @ 2.27GHz

CPUs compared: http://ark.intel.com/compare/75277,40201

kawsper3• 12 years ago

This is mine old $20 located in London, doing exactly nothing, I just created it last week, it seems very slow in comparison:

  dd bs=1M count=1024 if=/dev/zero of=test conv=fdatasync
  1024+0 records in
  1024+0 records out
  1073741824 bytes (1,1 GB) copied, 13,6791 s, 78,5 MB/s

  hdparm -tT /dev/xvda
  /dev/xvda:
  Timing cached reads:   3580 MB in  2.00 seconds = 1794.00 MB/sec
  Timing buffered disk reads: 336 MB in  3.01 seconds = 111.75 MB/sec

cbull• 12 years ago

Dallas data center, not upgraded, faster than London, slower than upgrade. ;)

  $> dd bs=1M count=1024 if=/dev/zero of=test conv=fdatasync
  1024+0 records in
  1024+0 records out
  1073741824 bytes (1.1 GB) copied, 2.19598 s, 489 MB/s

  $> hdparm -tT /dev/xvda
  /dev/xvda:
  Timing cached reads:   12446 MB in  1.98 seconds = 6273.29 MB/sec
  Timing buffered disk reads: 1768 MB in  3.00 seconds = 589.00 MB/sec

KamiNuvini• 12 years ago

Here's mine, with SSD upgrade:

  dd bs=1M count=1024 if=/dev/zero of=test conv=fdatasync
  1024+0 records in
  1024+0 records out
  1073741824 bytes (1.1 GB) copied, 2.17496 s, 494 MB/s

  hdparm -tT /dev/xvda
  /dev/xvda:
   Timing cached reads:   18560 MB in  1.98 seconds = 9358.78   MB/sec
   Timing buffered disk reads: 2510 MB in  3.00 seconds =   836.61 MB/sec`

growt• 12 years ago

Here is my $5 digitalocean slice (created a few months ago):

  # dd bs=1M count=1024 if=/dev/zero of=test conv=fdatasync
  1024+0 records in
  1024+0 records out
  1073741824 bytes (1.1 GB) copied, 3.7375 s, 287 MB/s

  # hdparm -tT /dev/disk/by-label/DOROOT

  /dev/disk/by-label/DOROOT:
  Timing cached reads:   16394 MB in  2.00 seconds = 8205.62 MB/sec
  Timing buffered disk reads: 868 MB in  3.00 seconds = 289.17 MB/sec

drzaiusapelord2• 12 years ago

287 MB/s vs 816 MB/s is huge. Linode is really doing a good job here. I have 4 VMs with them and am very happy. Can't wait to make this move.

saltcod2• 12 years ago

Just ran this on my 1GB droplet at Digital Ocean:

$ dd bs=1M count=1024 if=/dev/zero of=test conv=fdatasync 1024+0 records in 1024+0 records out 1073741824 bytes (1.1 GB) copied, 5.83509 s, 184 MB/s

saltcod2• 12 years ago

And here's a run I did on a shared host at A Small Orange for fun:

1024+0 records in 1024+0 records out 1073741824 bytes (1.1 GB) copied, 62.9023 s, 17.1 MB/s

Thaxll• 12 years ago

Those benchmarks are pointless on VM because of the memory caching...

kyrra3• 12 years ago

You're probably right. Do you think it's cache at the VM level before writing to disk. Would writing more data out help at all?

  $ dd bs=1M count=40960 if=/dev/zero of=test conv=fdatasync
  40960+0 records in
  40960+0 records out
  42949672960 bytes (43 GB) copied, 48.455 s, 886 MB/s

Nux• 12 years ago

Throughput speed is not everything, in many situations latency is more important. Can you give results for this?

dd if=/dev/zero of=test bs=512 count=1500 oflag=dsync

I'm getting "6.40264 s, 120 kB/s" on my laptop's SSD, but on a proper server with a RAID card and cache enabled I get "0.187033 s, 4.1 MB/s".

kbrower2• 12 years ago

dd if=/dev/zero of=test bs=512 count=1500 oflag=dsync 768000 bytes (768 kB) copied, 0.3325 s, 2.3 MB/s

kyrra3• 12 years ago

I did 7 runs of your test and the average time is: 0.424795s. Example run:

  $ dd if=/dev/zero of=test bs=512 count=1500 oflag=dsync
  1500+0 records in
  1500+0 records out
  768000 bytes (768 kB) copied, 0.427272 s, 1.8 MB/s

lloeki• 12 years ago

DO @5$/mo:

    $ dd if=/dev/zero of=test bs=512 count=1500 oflag=dsync         
    1500+0 records in
    1500+0 records out
    768000 bytes (768 kB) copied, 1.64471 s, 467 kB/s

diorray2• 12 years ago

DO @$40-AMS1

  1500+0 records in
  1500+0 records out
  768000 bytes (768 kB) copied, 1.17399 s, 654 kB/s

kbrower2• 12 years ago

Not a scientific benchmark, but I have a linode that has 8 cores before and after that just finished its upgrade. Previously it was consistently using just under 5/8 CPUs and is now using a ~2/8 CPUs with the same in/out traffic.

seunosewa• 12 years ago

How did you upgrade? I can't find any upgrade link, even after switching to a 64-bit kernel.

castis• 12 years ago

If your machines are eligible, there will be an upgrade button in the right nav when looking at one of your linodes.

jaequery5• 12 years ago

DO's biggest problem is their lack of "zero-downtime snapshot backup and upgrading". i've not used Linode but anyone know if theirs is any different?

SudoAlex2• 12 years ago

Upgrades/downgrades requires transferring all your data from one server to another - so that means downtime. The bigger the server, the more downtime in transferring.

As for backups, Linode backups seem to get a bad reputation. They do take a snapshot of a running instance - but reports from the forums indicate that they regularly fail, leaving you with a missing backup at some points. The only way of restoring a backup is to deploy that backup as a new server.

I'd recommend avoiding Linode backups and doing your own.

jaequery5• 12 years ago

upgrades for ram and storage does not have to have any downtime. same for snapshots, through LVM/ZFS, they can be nearly instant in most cases. it just seems like DO have not been able to figure this out for whatever reasons.

mnordhoff2• 12 years ago

"The only way of restoring a backup is to deploy that backup as a new server."

Not true. You can restore a backup to a node as long as it has enough unallocated disk space. The simplest ways to create enough unallocated disk space are to create a new node or to delete or resize all your disk images.

zak_mc_kracken2• 12 years ago

Does any of LINode or DigitalOcean offer plans without any SSD? I couldn't find any.

I just want to install some personal projects there for which even SSD's are overkill...

chc5• 12 years ago

Overkill never hurt anybody. Would you really want to ditch the SSD even if no other numbers changed at all? Even if you don't need it, I don't see why it's a bad thing.

zak_mc_kracken2• 12 years ago

It's just overkill to me to pay $20 a month for this when I expect a non-SSD plan would be half as cheap, if not more.

level09• 12 years ago

I would probably move back from Digital Ocean if they allow a 10$/mo plan.

I know that's not a big price difference, but some website really don't need a lot of resources. they work well on D.O's 5$ server, and I have really a lot of them.

munger2• 12 years ago

Rackspace cloud customer here… These Linode upgrades are very tempting to entice me to switch.

I get I might not be their target market (small business with about $1000/month on IaaS spending) but there are a couple things preventing me from doing so: 1) $10/month size suitable for a dev instance. 2) Some kind of scalable file storage solution with CDN integration – like RS CloudFiles/Akamai or AWS S3/Cloudfront or block storage to attach to an individual server.

I guess you get what you pay for… infrastructure components and flexibility AWS > RS > Linode > DO which roughly matches the price point.

beering• 12 years ago

I agree with your point about block storage - with Linode, there's a hard limit on how much hard drive space you get on any single server, which makes it a poor choice for anything that requires a lot of storage space (databases, file servers, etc.).

Unlike Amazon, you don't have any good options for increasing your storage space on Linode using network block storage, unless you somehow run another Linode and NFS them together. Being able to launch a moderately priced server with hundreds of gigs of spinning rust on AWS is still a nice option to have.

DougHaber• 12 years ago

I'm in a similar boat. I use Rackspace for their servers and CDN, and Linode is definitely starting to look nice. It could actually be cheaper to go with Linode and a separate CDN provider, such as MaxCDN. MaxCDN has some of the features that are missing from Cloud Files, which is real nice too. The main thing holding me back is the time it takes to migrate. I know they target the higher end, but hopefully RackSpace will readjust their prices as well at some point.

munger2• 12 years ago

Although, I guess there is nothing to stop me from running servers on Linode and continuing to use RS for Cloud Files/Akamai -- except for network speed of course. Linode can deploy to Dallas though, where I use RS so maybe the network speed won't be a big problem.

ff_• 12 years ago

Wow, that's beautiful. Currently I'm a DO customer (10$ plan), and if they had a 10$ plan I'd make the switch instantly.

madsushi• 12 years ago

Why do I pay Linode $20/month instead of paying DO $5/month(1)?

Because Linode treats their servers like kittens (upgrades, addons/options, support), and DO treats their servers like cattle. There's nothing wrong with the cattle model of managing servers. But I'm not using Chef or Puppet, I just have one server that I use to put stuff up on the internet and host a few services. And Linode treats that one solitary server better than any other VPS host in the world.

(1) I do have one DO box as a simple secondary DNS server, for provider redundancy

arb99• 12 years ago

DO?

erichmond• 12 years ago

Digital Ocean - https://www.digitalocean.com/

vbtechguy4• 12 years ago

yeah DO = DigitalOcean

mwcampbell3• 12 years ago

I don't see that distinction at all. Both Linode and DO provide persistent local storage. Both let you resize your server. And now, they're competitive on price; Linode even has hourly billing. Granted, DO has lower-end plans than Linode. But I don't see any indication that DO treats servers as disposable or fungible whereas Linode doesn't.

chc5• 12 years ago

I don't think anyone's ever gotten good hard numbers, but just from having seen a lot of people's anecdata over the years, my general impression is that Digital Ocean seems to have more short-term outages and to oversubscribe their boxes a bit more.

gtaylor6• 12 years ago

But I don't see any indication that DO treats servers as disposable or fungible whereas Linode doesn't.

I'm not sure "disposable" is a good word to describe DO's approach. We do know that they cram a lot more people per host machine, though. That may be where the "treat VMs like cattle" statement was getting at.

You need only fire up a few instances of comparable type, get some CPU load, and watch some basic metrics to determine this for yourself.

voidlogic2• 12 years ago

I just have one server that I use to put stuff up on the internet and host a few services. And Linode treats that one solitary server better than any other VPS host in the world.

I use DO for this and have used many others in the past and I have never had any problems with DO...

gtaylor6• 12 years ago

Anecdotally, I get notifications of maintenance outages every month with DO in NYC1 and NYC2. These are exceedingly, exceedingly rare with Linode's Atlanta DC.

Also, stolen CPU on my DO boxes is super high, whereas this is very rare on Linode. You have less neighbors.

teekert• 12 years ago

Last month DO rebooted my droplet (I received an email notice of this), it came back with all my config files filled with just @@@@'s... They were not helpful at all, after 5 emails they offered to help migrate leftover data to a new droplet but there was nothing that wasn't in my backups, my site was down for a day.

vbtechguy4• 12 years ago

Not exactly a fair comparison, you'd want to compare similar plans, 2GB Linode VPS and 2GB DigitalOcean VPS are both $20/month. Some benchmarks between the two posted at https://blog.centminmod.com/346. Linode fast disk i/o than DO but DO faster cpu than Linode :)

novaleaf• 12 years ago

i'm a happy DO user (30 droplets at the moment) and I agree with your cattle analogy. I am not surprised to see multiple droplets spontaneously rebooting on occasion (even when system maintenance has not been announced)

however, I have no experience with Linode so I can not say if they are "kittens" as you say.

edwinyzh2• 12 years ago

Just a side node, I occasionally use SSH tunneling to get through the evil local government Internet censoring, so I have a linode a DO (just sever months), and linode's SSH tunneling is much much more fast and stable.

pangram• 12 years ago

That could be caused by the routing, as well, though -- are the servers in the same approximate geographical location?

mike-cardwell• 12 years ago

I moved from Linode to Digital Ocean because of the price and I've noticed no change in level of service. And the support, which I've used several times at both has always been incredibly quick, accurate and useful.

I was however happy to have a viable alternative to Linode. I found their handling of the two hacking incidents to be completely disrespectful to me as a customer.

rogerdpack• 12 years ago

Unfortunately, DO vs. Linode, DO is the only one that can provide me a "reasonably fast box" [cpu, ssd] for $5/month, so I'm cattle, but cheap... :) [pingdom shows 5min worth of outage in the past week (a 99.96% uptime) which may or may not be OK depending on your money vs. uptime pain threshold :) ]

dharma1• 12 years ago

ohhh yesss. DO is good for some locations like Southeast Asia but loving this upgrade for my London and Tokyo Linodes

notastartup4• 12 years ago

These upgrades are impressive but they are a bit too late to the game. DO still has these advantages besides the cheap monthly price:

- DO has excellent and easy to understand API - Step by step guides on setting up and running anything - Minimal and simple

To entice me, it's no longer just a matter of price, DO has extra value added, largely due to their simplicity.

jsmthrowaway3• 12 years ago

You mean like linode.com/api/ and library.linode.com?

notastartup4• 12 years ago

yeah it's not as obvious and simple like DO's REST calls. shame I never discovered this while I Was with linode but it wouldn't have caught my attention any how like how DO did with their simplicity.

now I definitely don't think there's much linode can do further to entice me unless they matched digital ocean's pricing lower than 0.03/hr

vidyesh3• 12 years ago

So this makes Lindode practically on par with DO's $20 plan. Up till now $20 plan at DO was better now its just the choice of the brand.

But here is one thing that DO provides and I think Linode too should, you get the choice to spin up a $5 instance anytime in your account for any small project or a test instance which you cannot on Linode.

yogo• 12 years ago

Doesn't hourly billing address that issue? Linode now offers it.

damian20002• 12 years ago

Yeah, that's the impression I got with hourly billing as well.

orthecreedence• 12 years ago

Bummer, they're taking away 8 cores for the cheap plans and replacing it with 2. Does anyone know if the new processors will offset this difference? I don't know the specs of the processors.

Linode's announcements usually come in triples...I'm excited for number three. Let's hope its some kind of cheap storage service.

opendais3• 12 years ago

tl;dr: If you had 40 VMs on 8 cores before, you have 10 on 2 cores now. It is the same ratio of VMs:Cores but with stronger processors.

Long version: "If you take the upgrade, you inherit the new plan specs, vcpus and all.

From Caker's comment on the blog. It seems that this was done to reduce fighting over core and provided more consistent fair availability of processing power when they tested it b/t VMs.

drzaiusapelord2• 12 years ago

Cores presented to your machine by the hypervisor do not have a 1 to 1 relationship with physical cores. In the world of virtual machines, how many cores you have is meaningless.

For example, my VMWare server is a single 8 core processor, but my VM's only see 2 cores, as that's the way I prefer it and I believe VMWare recommends this or even one core. Those 2 virtual cores can access all 8 physical cores.

The only real way to find out your cpu performance is to run a benchmark.

personZ• 12 years ago

If you had a host machine with 8 cores, and a virtual machine with 2 cores, the virtual machine will only possibly use 25% of the CPU resources of the host machine -- one vCPU cannot be schedule in parallel across more than one physical core.

However it is advisable to minimize the vCPUs in VMs because it greatly enhances scheduling -- if a VM has 8 vCPUs allocated to it, with some hypervisors that lack advanced co-scheduled it won't be scheduled until 8 cores are available, and in some situations it will sit on 8 cores during its duration whether it is only using 1 or all. If you have a large number of smaller VMs, scheduling can greatly improves.

This isn't true for all hypervisors and situations. You mileage may vary.

damian20002• 12 years ago

That's the question I had too, thanks.

conorh• 12 years ago

Benchmarking using wrk the smallest linode (1024 now 2048) serving a page from an untuned Rails application using nginx/passenger getting almost no other traffic. Hard to compare of course given the various other factors, but produced slightly lower performance after the upgrade. Serving a page from nginx directly (no Rails) had no appreciable difference in performance, I guess the Rails web serving is more vCPU bound?

Before Upgrade:

  Running 30s test @ http://...
    5 threads and 20 connections
    Thread Stats   Avg      Stdev     Max   +/- Stdev
      Latency   308.91ms  135.01ms 985.82ms   80.00%
      Req/Sec    14.15      4.61    24.00     66.36%
    2206 requests in 30.00s, 28.51MB read
  Requests/sec:     73.53
  Transfer/sec:      0.95MB

After Upgrade:

  Running 30s test @ http://..
    5 threads and 20 connections
    Thread Stats   Avg      Stdev     Max   +/- Stdev
      Latency   321.74ms  102.45ms 957.74ms   87.32%
      Req/Sec    12.02      2.18    17.00     80.75%
    1858 requests in 30.01s, 24.03MB read
  Requests/sec:     61.92
  Transfer/sec:    819.98KB

chc5• 12 years ago

It depends on the task. If your Rails app hits the database a lot, memory availability and disk speed will matter a lot, which favors the new Linode style. If your app just does one small query per page and then gets on with its work, then probably CPU is your most important thing (since most of your time will be spent on app logic, runtime housekeeping and rendering templates).

cmelbye• 12 years ago

Yeah, you went from 8 vCPU Cores to 2 vCPU Cores on that plan as part of the upgrade. It seems the processor upgrade makes up for that decrease a little bit, but it is a decrease.

ihowlatthemoon• 12 years ago

VPSBench result:

Before

-------

  CPU model:  Intel(R) Xeon(R) CPU           L5520  @ 2.27GHz
  Number of cores: 8
  CPU frequency:  2266.788 MHz
  Total amount of RAM: 988 MB
  Total amount of swap: 255 MB
  System uptime:   8 days, 12:03,
  I/O speed:  69.9 MB/s
  Bzip 25MB: 8.96s
  Download 100MB file: 47.2MB/s

After

------

  CPU model:  Intel(R) Xeon(R) CPU E5-2680 v2 @ 2.80GHz
  Number of cores: 2
  CPU frequency:  2800.086 MHz
  Total amount of RAM: 1993 MB
  Total amount of swap: 255 MB
  System uptime:   2 min,
  I/O speed:  638 MB/s
  Bzip 25MB: 5.10s
  Download 100MB file: 146MB/s

Test: https://github.com/mgutz/vpsbench

pdq• 12 years ago

Here is DigitalOcean's $5/month plan:

  CPU model:  QEMU Virtual CPU version 1.0
  Number of cores: 1
  CPU frequency:  2299.998 MHz
  Total amount of RAM: 491 MB
  Total amount of swap: 0 MB
  System uptime:   276 days, 5:04,       
  I/O speed:  253 MB/s
  Bzip 25MB: 7.83s
  Download 100MB file: 56.0MB/s

Here is Vultr's $5/month plan:

  CPU model:  Vultr Virtual CPU 2
  Number of cores: 1
  CPU frequency:  3399.994 MHz
  Total amount of RAM: 490 MB
  Total amount of swap: 509 MB
  System uptime:   28 days, 3:15,       
  I/O speed:  485 MB/s
  Bzip 25MB: 3.17s
  Download 100MB file: 58.2MB/s

harrystone• 12 years ago

I would love to see them still keep all those old disks and sell me some huge, cheap, and slow storage on them.

pavanky2• 12 years ago

I wish Linode (or anyone else other than Amazon) provides a reasonable Plan[1] with GPUs on them.

[1]: Amazon charges $2 an hour thats about $1500 a month.

dugmartin• 12 years ago

Here is a list with five other companies:

http://www.nvidia.com/object/gpu-cloud-computing-services.ht...

pavanky2• 12 years ago

Thanks for pointing it out. I should have known better considering that I am in the space :(

thenomad• 12 years ago

Interesting - thanks! Are any of those plans more reasonably priced than Amazon's?

ldargin• 12 years ago

Hmm, Amazon now charges $0.65 per hour for their g2.2xlarge GPU plan (in Linux). See http://aws.amazon.com/ec2/pricing/

gatehouse• 12 years ago

I checked spot instances just now, and it was less than $0.10 an hour. So for batch jobs that might be an option. And if you're using it interactively, than you can always go EBS backed and just turn it off when you're done.

hyptos• 12 years ago

wow EC2 instance free plan :

$ dd bs=1M count=1024 if=/dev/zero of=test conv=fdatasync

1024+0 records in

1024+0 records out

1073741824 bytes (1.1 GB) copied, 35.8268 s, 30.0 MB/s

vacri• 12 years ago

Yeah, AWS doesn't have the best speed to drive, though if you have money, you can specify how many IOPS you want... you just have to pay for your declared use every hour whether you're using it or not. It's also kinda hard to predict if you're not familiar with that kind of thing.

Note that on AWS your hard drive isn't on the machine running your VM. If you want to test local disk speed, do it in /mnt, where local disk storage is mounted... but remember that data stored there won't persist if you turn your instance off.

shiloa2• 12 years ago

I have mixed feelings about this. We're in the process of moving from Linode to Rackspace but haven't flicked the switch just yet - was planning to this weekend.

Our Linode server (16 GB plan) has been performing terrible lately wrt I/O (compared to, say, a Macbook Pro running the same computations), and we decided we've had enough. I guess we'll have to compare the two after the upgrade and decide.

thraxil• 12 years ago

Did you ever put in a support ticket? I've had I/O performance issues with Linode boxes before and it's always been quickly fixed when they are notified.

shiloa2• 12 years ago

I did, they switched us to a new machine - no change. Their support was great, but the hardware needed to catch up. Perhaps this upgrade will help.

beedogs• 12 years ago

This is nice to see. SSD has gotten ridiculously cheap lately.

jrockway• 12 years ago

A nice reward for those of us who have been using Linode from before they even had x86_64 images.

simcop2387• 12 years ago

Yes. I'm actually looking at upgrading my server to a 64bit kernel just so i can take advantage of this. I've been on them so long i'm still a 32bit userland.

filmgirlcw3• 12 years ago

Shall we call this the DigialOcean effect?

weixiyen2• 12 years ago

sigh... just put 12 droplets in digitalocean last week. now thinking of switching back...

jdprgm• 12 years ago

This is really a fantastic upgrade. I've been hosting with Linode for a few months now and been very happy with them. I run a relatively transfer intense SaaS app and a 50% transfer increase makes quite an improvement.

rdl• 12 years ago

Semi-related: does anyone know of any good (but still fairly cheap) providers doing Atom C2750/C2758 servers yet?

__xtrimsky3• 12 years ago

I still prefer OVH.com http://www.ovh.com/us/vps/vps-classic.xml

for $7 you get: 2 cores 2GB RAM

for 10$ you get: 3 cores 4GB RAM

They don't have SSD, but SSD doesn't do everything, I prefer more ram.

EDIT: If some of you don't know OVH, it's because its new in America, but its not some cheap company, it's a European company that is very successful there. And just recently created a datacenter in North America. (I used to live in France, and have known them for some years).

giulianob5• 12 years ago

Can you run https://github.com/mgutz/vpsbench and see how the performance is to compare?

__xtrimsky3• 12 years ago

I have the 7$ VPS, here is the result:

  _04/17/2014 - VMPLAN - DATACENTER - OS - AUTHOR_
  ```
  CPU model:  AMD Opteron(tm) Processor 4284
  Number of cores: 2
  CPU frequency:  3000.000 MHz
  Total amount of RAM: 2048 MB
  Total amount of swap: 128 MB
  System uptime:   14 days, 13:31,
  I/O speed:  75.4 MB/s
  Bzip 25MB: 7.44s
  Download 100MB file: 13.0MB/s
  ```

gingerlime• 12 years ago

vpsbench on a 2Gb fresh Linode in Dallas

  ```
  CPU model:  Intel(R) Xeon(R) CPU E5-2680 v2 @ 2.80GHz
  Number of cores: 2
  CPU frequency:  2800.050 MHz
  Total amount of RAM: 1993 MB
  Total amount of swap: 511 MB
  System uptime:   0 min,       
  I/O speed:  811 MB/s
  Bzip 25MB: 4.93s
  Download 100MB file: 19.1MB/s
  ```

abronte• 12 years ago

I have the 4gb vmware vps (http://www.ovh.com/us/vps/vps-cloud.xml)

  _04/17/2014 - VMPLAN - DATACENTER - OS - AUTHOR_
  ```
  CPU model:  AMD Opteron(tm) Processor 4386
  Number of cores: 4
  CPU frequency:  3100.000 MHz
  Total amount of RAM: 3943 MB
  Total amount of swap: 1998 MB
  System uptime:   10 days, 2:25,
  I/O speed:  169 MB/s
  Bzip 25MB: 5.34s
  Download 100MB file: 15.8MB/s
  ```

mappu• 12 years ago

Neat script! Here are some low-end results for comparison.

BuyVM 128MB ram, 15GB disk, 15USD/year

  CPU model:  Intel(R) Xeon(R) CPU L5520 @ 2.27GHz
  Number of cores: 1
  CPU frequency:  2266.746 MHz
  Total amount of RAM: 128 MB
  Total amount of swap: 128 MB
  System uptime:   39 days, 22:35,
  I/O speed:  89.9 MB/s
  Bzip 25MB: 12.89s
  Download 100MB file: 50.0MB/s

QuickPacket: 128MB ram, 20GB disk, 15USD/year

  CPU model:  Intel(R) Xeon(R) CPU X3470 @ 2.93GHz
  Number of cores: 1
  CPU frequency:  2933.228 MHz
  Total amount of RAM: 128 MB
  Total amount of swap: 128 MB
  System uptime:   48 days, 6:07,
  I/O speed:  129 MB/s
  Bzip 25MB: 6.40s
  Download 100MB file: 1.14MB/s

A good aggregator for low-end deals is lowendstock.com .

kh_hk2• 12 years ago

Speaking of OVH, their catalog on low-end dedicated servers (when available) is quite nice too http://www.kimsufi.com

hadoukenio• 12 years ago

Just be aware that OVH is using OpenVZ whereas both Linode and DO are using KVM. That's a big difference if it matters to you

porker3• 12 years ago

I don't understand the difference between OpenVZ, KVM and all the other virtualization providers. Is it that some are much faster than others?

hanru• 12 years ago

Linode is Xen; DO is KVM.

bluedino3• 12 years ago

OVH has two VPS offerings: 'Classic' which runs OpenVZ and 'cloud' which runs on VMware, and also is priced higher

ausjke2• 12 years ago

This is great indeed. I'm happy Linode did this. I ran below command 10 times and used the average below:

dd bs=1M count=1024 if=/dev/zero of=test conv=fdatasync

Linode: 1073741824 bytes (1.1 GB) copied, 1.09063 s, 985 MB/s D.O: 1073741824 bytes (1.1 GB) copied, 3.23998 s, 331 MB/s

dd if=/dev/zero of=test bs=512 count=1500 oflag=dsync

Linode: 768000 bytes (768 kB) copied, 0.478633 s, 1.6 MB/s D.O: 768000 bytes (768 kB) copied, 1.01716 s, 755 kB/s

nilved2• 12 years ago

Linode's recent upgrades are awesome, but people are very quick to forget the period where they were being hacked left and right and didn't communicate with their customers until a defensive blog post weeks after the fact. No matter how good the servers may be, Linode should be a non-starter for anybody who cares about the security of their droplet; and, if you don't, why would you pay Linode's premium fee?

jsmthrowaway3• 12 years ago

Ask DigitalOcean if they're wiping data on droplet cancellation and snapshot deletion yet. They've fucked that one up several times since the inception of the company and each time get a little more defensive about it. For a while there, I could spin up a droplet and get database passwords, keys, data, all day long from the previous owner. Then they told the guy that reported it that he was mistaken, and basically lied in a blog post and said no user data was at risk.

I'm told that's just the beginning of security faults in their platform, attributable to being a younger company and discovering these things for the first time; I have been personally shown evidence that DigitalOcean's platform initially trusted a hidden field called "userid" and allowed a user to operate on any other user without authorization, including restoring images, shutting down droplets, and so on. Their system at first had no protection against spoofed packets exiting a droplet, either, so ARP poisoning the gateway was (and possibly remains) a viable attack.

Linode is far more mature, obviously.

Every provider has security issues. It's how they rectify and move forward that should concern you. Watching DigitalOcean react to being informed that the issue they got burned on once had reappeared basically told me to never use their services.

Also, and I actually consider this very important and not a grammatical nit, Linode doesn't sell droplets. They sell virtual servers. You're asking for a Pepsi from Coca-Cola. You might consider this a minor nit, but it's actually a serious confusion issue that I already see happening.

Allow me to promise you -- not predict, promise -- that DigitalOcean will be compromised just as badly. It's going to happen. It's a matter of when and how they react.

nilved2• 12 years ago

I don't see why your (agreeable) sentiment about Digital Ocean affects the fact that Linode is bad at security and customer interaction. We can always avoid both.

I didn't mean to use the word "droplet," so I'm sorry for the confusion.

bambam128972• 12 years ago

It's amazing how short people's memories are. They behaved in a very scummy way back when that all happened. I was sure they would have to close their doors within 6 month.

But it's exactly one year since the hack and everyone is back to talking about how amazing Linode is

http://news.softpedia.com/news/Hackers-Claim-to-Have-Gained-...

__xtrimsky3• 12 years ago

Could someone please explain what improvements can we get from SSD for web applications ?

I know it would read files faster, but in most cases reading a couple of PHP files is not such a big improvement.

My guess would be maybe databases ? Read time improvement for MySQL ?

Kudos• 12 years ago

Ubuntu 14.04 LTS is now available on Linode too.

corford• 12 years ago

Big shame the new $20 plan now only offers 2 cores versus 8 with the current plan. For my workloads, I don't need 2GB RAM or SSD disks, I just need the cores :(

aroman• 12 years ago

Interestingly, I had the opposite reaction. I plan on running node.js applications on Linode, and as node cannot automatically take advantage of multiple cores (and even manually the support is experimental), the announcement of vertically scaled but horizontally contracted CPU cores was very exciting to me.

SCdF• 12 years ago

Linodes are now SSD. This is not a hybrid solution – it’s fully native SSD servers using battery-backed hardware RAID. No spinning rust! And, no consumer SSDs either – we’re using only reliable, insanely fast, datacenter-grade SSDs that won’t slow down over time. These suckers are not cheap.

http://techreport.com/review/26058/the-ssd-endurance-experim...

Not to slam what Linode is doing here, and I'm sure there are probably lots of great reasons to buy datacentre-grade SSDs, but just thought I'd point out that slowing down over time (or data integrity issues) are not really consumer-grade problems any more :-)

SiVal• 12 years ago

What has changed? (Not advocacy; I'm curious.)

kolev2• 12 years ago

Goodbye, Digital Ocean!

kolev2• 12 years ago

You can downvote as much as you want, but having faster nodes from a long-established company, which doesn't keep your old images available to new customers to peek in, beats DO! I've been with DO since they launched and I can't say they are one of the more reliable services I've dealt with!

funkyy• 12 years ago

I would love to see Linode going to large HDD drives option for storage as well. I am dying to find really inexpensive cloud provider with cheap data space (SATA is fine), reasonable bandwidth but low cpu and ram and Linode style support/caring. Give server with ~500 GB hard drive, 2 TB outgoing transfer, 1 core and 1 GB ram for ~$20-30 and I am all yours.

kijin• 12 years ago

About a week ago, I wrote a comment in another Linode-related thread asking how the new usage patterns that hourly billing encourages might affect CPU contention. At the time, I received 11 upvotes but no replies. Apparently, quite a few people were interested in my question but had no useful conjectures to share.

https://news.ycombinator.com/item?id=7564764

Now it's obvious what Linode's answer to that question is: Lower "burstable" CPU for lower plans.

The $20 plan used to be able to burst to 8 cores for short periods, but now it only has access to 2 vcores. The "guaranteed" processing power is probably higher with the newer CPUs, but at the expense of short-term burst performance.

Another minor detail that I find interesting is that the transfer cap for the $20 plan has been increased to 3TB, whereas the $40 plan still gets 4TB. Apart from the transfer cap plateau-ing at the extreme high end, this is the first time that Linode has broken its 11-year-old policy of "pay X times as much money, get X times as much RAM/disk/transfer".

jebblue• 12 years ago

I was looking into alternatives but now I'll stick with them, I can't find another cloud provider whose stuff works so well.

edit: I just finished the migration, my disk speed test is through the roof, free ram is phenomenal!

EGreg• 12 years ago

I love linode. I switched from slicehost for its 32bitness back in the day, stayed for the awesome culture and independence. Slice host got sold to rack space.

However, I am seriously considering a move to Amazon Web services for one main reason: I need to decouple the hard drive space from the ram. The hard drive space is so expensive on linodes!!

ksec• 12 years ago

Sometimes i just wish the pricing system would get better as you go larger.

What is the difference between the 16GB - 96GB Plan and a dedicated server? And why would i pay 3x the price? The advantage of those who offer Cloud / VPS and Dedicated Servers Hosting company is they can mix and match depending usage. If you are actually building an any sort of infrastructure with Linode those large box are extremely expensive.

icantthinkofone• 12 years ago

Without FreeBSD support, it means nothing to me.

vbtechguy4• 12 years ago

Updated benchmark results with 2GB vs 4GB vs 8GB vs 16gb plans from Linode vs DigitalOcean https://blog.centminmod.com/346. Definitely Linode has the faster cpus and disk i/o as you move up in plans >2GB. 16GB plans are pretty close though if you look at subtests in UnixBench and ignore the subtests affected by different base Linux Kernel versions used.

mark_lee• 12 years ago

awesome, linode or DO, if you're small or media companies, no other options should be considered at all, even AWS or Google Cloud.

ForFreedom• 12 years ago

The specs for $20 is same for DO and Linode excepting for the 8GB extra HDD on Linde.

The New Linode Cloud: SSDs, Double RAM and much more

Related Stories