After all the talk about how much cheaper dedicated hosting can be compared to AWS, I looked into dedicated hosting recently. It didn't seem as great as the comments here led me to believe. Rackspace for example starts at $499/month for a server with 4GB of RAM. On AWS I can get a c1.xlarge instance with 7GB of RAM for $435/month, and that's with paying the on-demand price 24x7.
Ok, that doesn't include EBS and bandwidth. So let's throw in 1TB of outbound data (less than the 2TB that Rackspace includes, but still more than I'd ever use in a month) and a 150GB EBS volume with 1000 provision IOPS. Let's also use the 1 year heavy utilization price while we're at it. Now I'm looking $342/month with $1800 upfront, which amortizes to $492/month.
With AWS I also don't need to stick with paying 24x7 for infrastructure. I don't need to keep my CI infrastructure running on nights and weekends for example, and with AWS I have the flexibility to turn that into cost savings.
So I don't see the order of magnitude price difference. So maybe I shouldn't be looking at Rackspace? Hetzner keeps coming up on here, but I don't need a server in Europe. Maybe OVH? Hmm, the comments on Hacker News don't exactly inspire confidence in OVH for something more than a personal server to screw around with. I need to run infrastructure for my business. Do I start hunting for bargain hosting providers on WebHostingTalk?
Ok, back to AWS it is for me.
EDIT: Thanks for the suggestions to look at Softlayer. A server with 1 CPU and 8GB of RAM starts at $399, which is better than Rackspace, but not exactly orders of magnitude cheaper than the c1.xlarge I mentioned above.
I think you picked the most expensive dedicated hosting provider in existence. Have a look at SoftLayer, Hetzner, Incero, or one of the other alternatives. Also, at these prices, it's not unreasonable to just order any old Dell server for $600 and put it into a colo rack somewhere.
For reference, the dedicated server I rent for my private projects comes in at €49/month and it has these specs:
Intel® Core™ i7-2600 Quad-Core
16 GB DDR 3 RAM
2 x 3 TB SATA 6 Gb/s HDD; 7200 rpm
Support is extremely fast and competent, I'm really happy. To try and mirror that package with any kind of Amazon (or Rackspace it seems) offering would be prohibitively costly.
No ECC (http://en.wikipedia.org/wiki/ECC_memory) in that configuration. That's a serious no no for running anything but personal/low importance stuff. Random bit flips do happen and will ruin your day.
This is highly unlikely. I'm on a team that administers an EC2 infrastructure with tens of terabytes of RAM, and I've never seen anything that resembles memory errors. Considering the findings of http://www.cs.toronto.edu/~bianca/papers/sigmetrics09.pdf, this would be rather extraordinary.
The whole point of the "cloud" is to use as commodity hardware as possible... accepting that failures happen. If you aren't using multiple regions and servers for deployment, with failover designs in your applications you are doing it wrong.
Run n times, compare results. The idea of the same bit flipping twice is fairly ludicrous. These n times could be on the same machine, or distributed and compared with paxos. Then back everything up a billion times. Safety is extremely hard to guarantee, so the closest approximations will be expensive.
I doubt that would cause an instance to fail entirely in a majority of instances. Linux has it's own memory corruption checking and fixing (on 64-bit words afaik), also the customer can code their own sanity checking or CRC data to detect and eliminate them as much as possible. Also, many file or copy operations have built in corruption checking as well. Of course, on top of that you have your virtualization platform which I'm sure has it's own series of sanity checks and complex hardware handling logic.
Ultimately, it's a business decision of risk-management. Also, if you're in any kind of serious business like banking, there is already regulations around the standards of hardware/software you can use.
Users will never know because EC2 runs on paravirtualized kernels that don't have access to SMBIOS, which would tell us that information.
However, if we can determine what machine models they use (assuming they buy branded hardware), then we will know their chipsets, and from there, we should be able to deduce whether the RAM is ECC or not.
Actually when I was at NetApp we got pretty good at figuring out ECC failures. Just write a program that allocates 1GB array. Fill it with random data and compute a MD5 hash on it (you can do that fairly quickly). Now recompute the hash once every 5 minutes for a month. Run that on 125 instances. (that ends up being 1 Terabit of data btw) which makes a good detector for cosmic rays if it isn't ECC protected :-).
I did look at SoftLayer and it's still not significantly cheaper than AWS. Hetzner doesn't seem to have servers in the US, so I've already ruled them out.
Incero seems cheap, but that's the only thing I've seen about Incero. I consider Incero to be in the bargain hunting on WebHostingTalk category of providers. I'm not entirely comfortable hosting mission critical infrastructure there yet.
There are probably a lot of alternatives that other people know more about than I do. Just to address a concern with Hetzner: connectivity is really good there. Last time I was in the US, I got better roundtrip times from my German server than I got from MediaTemple gs in California (though they also have some very nice features). Of course there is no going around light speed the limitation, but I'd say for a lot of projects a European server +CloudFlare is totally sufficient. In the end, it's all a tradeoff between different factors ;)
An issue with virtually all dedicated server vendors is the tendency to grossly under-equip servers with memory: My Mac Mini has 16GB (a $100 upgrade), and it is simply ridiculous that going above 4GB on a server is considered some extravagant feature the requires significant, expensive upgrades of every other component.
In VPS, memory is everything. Memory is the # of VMs they can spin up, which is directly proportional to the amount of money they're making. What you're ultimately paying for is memory. Yes, some bandwidth, but mostly memory. It's what hotel rooms are for hotels.
If we were talking about VPS servers, then sure, however we're talking about dedicated servers. What you are renting is a physical box (such as Dell R210s) sitting in a rack somewhere.
Is it possible that they might want to save money by giving you wimpy memory, because they feel that with the 32gb they give you, they could be making a lot more with something like 64 micros?
16GB of ECC memory for servers is a lot more expensive than your Mac Mini's memory. I recently spent £250 on 2x 4GB DIMMs for a Dell Poweredge server[1] (ok DDR2 is older and therefore more expensive, but still). The same configuration for my Mac Mini came in at around £50[2]
You paid a significant premium for that older memory (a funny paradox of the industry). As P1esk mentioned, you could have gotten 32GB for around the same price. Straight from Dell, paying their inflated prices, on a new R210 it's $173 to go to 16GB of ECC memory.
I've bought a number of large servers recently (using owned servers at colocation in combination with virtual hosts for DR and geo-sharding), and always equip them with 192GB or more. The pricing of memory is so incredibly low it makes no sense otherwise, yet all of these dedicated server vendors act as if 2GB is the norm and 4GB is the advanced upgrade.
Also checkout M5 hosting[1][2]. They have excellent customer support and I've had no issues since I started using them a few months ago. I currently have a dedicated freebsd server, for $209/month:
Xeon X3450 Quad Core, 32GB ECC RAM, 2x 500GB HD, and 2TB/month bandwidth plus a whole bunch of extra IPs allowing me to run numerous jails for hosting all sorts of sites and dev-environments. Very good machine for the price. You can still pick one up [3].
If cash cost is your only consideration it might make sense to look outside of AWS. One of the big issues I see is that the feature-set and global footprint of AWS is pretty much unmatched.
The really sad part (for AWS competitors) is that once you get into reserved instance pricing and spot pricing, it removes a lot of the cost difference. The competition is years behind, and getting further behind every month.
But that's looking at it from just a cash perspective. If your time is worth money (and you have funding), setting up things on AWS will save you a ton of money since you can take advantage of pretty good services like RDS, ELB, Elasticache, SQS, Route53, etc that will allow you to get your product to market faster with a smaller team.
To be honest, you chose looking at the most expensive dedicated hosting provider. Also, they're just not-that-great. Their service and support staff are wonderful, but as soon as you actually need to do anything complex... good luck getting an engineer on the phone. It used to take us three or more calls to get someone higher level on the phone and we were paying five-figure type sums to them each month.
Also, we got DoS attacks all the time on their network and not one since we moved to our own hardware... Obviously take this with a grain of salt, but take the other comments to heart and actually look into other dedicated provider pricing. You might be surprised.
Looking from Rackspace to OVH is really going straight from one extreme to another. Softlayer is a more sensible middle-ground (though note their list prices are high, you'll almost always be able to get 2x RAM or similar).
Thank you for the recommendation, it's definitely a better deal than Rackspace for my needs, but not compelling enough for me to switch from AWS at the moment.
If you need raw performance, have a look at WebNX, they are specialized in high end dedicated servers, boxes with Xeon E5 and 64GB RAM are in the range $150-$200. I find Softlayer or Rackspace quite overpriced.
There are many other smaller hosting companies that offer good services and pricing. You just need to do some researchs.
Don't forget to include all the free bandwidth that comes with dedicated hosting plans. With EC2, that's extra and can really make a difference to a lot of companies.
And, with dedicated hardware, you have faster storage with traditional HDD and even SSD. Both those options are faster than EC2 storage.
The startup I'm at has looked at using AWS and repeatedly decided against it based on cost. It's simply much more expensive compared to dedicated hardware when you need a single high-performance box -- and you should be able to serve a whole fuckton of traffic an 8-cpu 32GB box if you try.
4 CPU and 7GB ram on Windows Azure is $268/month if you use Windows, and $179/month if you are using Linux. Add $120 for 1TB outbound traffic, and you are looking at $388/month Windows and $299 for Linux.
IBM closed on Softlayer a few days ago. Not many people expect Softlayer to continue thriving the way it did over the years. But at the rate Softlayer grew it was apparent they were eyeing a quick exit (not too common in the dedi business).
> EDIT: Thanks for the suggestions to look at Softlayer. A server with 1 CPU and 8GB of RAM starts at $399, which is better than Rackspace, but not exactly orders of magnitude cheaper than the c1.xlarge I mentioned above.
You need to negotiate. The advertised prices are starting points for talking to sales, not what you really have to pay.
You can get a 2011-gen quad-core Xeon with 16GB RAM for $250/mo or thereabouts at Softlayer.
^ An important point. I emailed SoftLayer sales when setting up Alikewise and got a much better price, around $300/mo.
This is probably still overpaying for my needs, but substantially cheaper than buying via their web interface.
Another thing I haven’t done is renegotiate in the last 2 years. Moore’s law and all, I suspect I wouldn’t save $$ but would get more horespower for the same price.
Yeah, I've already ruled out OVH if you read the last part of my comment. The comments on Hacker News about OVH didn't really inspire confidence in OVH for more than hosting side projects, and that's not my use case.
Are you using OVH for hosting Exocortex? I checked your website and it seems to be hosted on Linode.
OVH is not bad, most people that they cut out are people who are new, they rent 10 servers or more, start spam, then they get cut and they go on forums to complains. Since OVH is a lower cost provider, they attrack more customer like this.
On the other end, Softlayer is very reliable, I have 4 servers with them and I had nearly no problems, they have the best network as well (I got fewer network outage with Softlayer then Rackspace).
Rackspace is better if you need something managed, but they are expensive.
in case you are unaware, first, see the perf difference in my previous post, and second, the public pricing on softlayer is a price ceiling. If you are buying multiple machines, even perhaps as few as 4 or 5 a month, you can probably negotiate a nice discount and certainly if you are renting 10-20/mo. Of course they will happily sell for the price on the site.
Same here. These "amazon is so expensive and so slow" posts show up here constantly, and every time I try to get information on alternatives, and nobody has anything. A single dedicated server running in some warehouse is not a replacement for AWS. I need S3, or I need a bunch of storage devices in multiple datacenters automatically keeping the files I put on them synced. Which hosting provider is doing that again? Where is my EBS with a dedicated server warehouse? I need block device that is durable, and can be quickly and easily detached from one server and attached to another. Even the cheapest redundant iSCSI SAN setup is very expensive.
I am not paying for AWS because I am dumb, I am paying for AWS because they provide services that allow me to outsource the high durability to them. Someone show me an actual alternative please, rather than just repeating "zomghosting.biz gives you a million gigs for $20!!1".
...and the people who criticize AWS aren't dumb, either. AWS brings a lot to the table, and also takes a few things off the table. AWS's scaling and combo of services are both wonderful, but they come at the cost of performance and price.
Most companies don't need "click of a button"-type scaling of instances. And most people don't need all of AWS' services; replacing them with good, stable open-source software alternatives (eg., RabbitMQ for SQS, Riak for Dynamo, Postgres for RDS), while incurring a higher admin cost, is not non-trivial.
My company replaced all of our AWS instances with co-located servers a couple of years ago. Our servers were basically static; we did not rely on any dynamic scaling of instances, and since we are in Europe, we couldn't rely on datacenter distribution, either. The performance difference was extreme. The servers were cheap, but they still blew Amazon out of the water. We could easily afford two 32-core database servers within our budget, and their performance translates to a whole bunch of Amazon's fastest instances.
The up-front hardware costs were less than what we were paying per year to Amazon. We have an entire rack now, and I think we are paying something like $1,500/mo for the entire hosting package. (It may be a little more, but not much.)
Hardware maintenance costs have been negligible so far, and the redundancy built into our infrastructure and stack means we don't have to rush out the door if a box dies on us.
>and the people who criticize AWS aren't dumb, either
I didn't say they were. I asked them to provide an alternative rather than pretending a single dedicated server is comparable to AWS.
>Most companies
I don't care what most companies need. I am not trying to argue, I am asking the people insisting that nobody should ever use amazon to provide a reasonable alternative or stop making sweeping generalizations.
>We could easily afford two 32-core database servers within our budget, and their performance translates to a whole bunch of Amazon's fastest instances.
Did you seriously not notice how you aren't addressing anything I said, but just repeating exactly what I pointed out wasn't relevant? I don't need more cores, I need reliable storage. What is my EBS? Until we're big enough to be running our own datacenters, there's not really a reasonable alternative that I can find. I am asking for advice, not "more cores!".
Cloud Sigma (http://www.cloudsigma.com/) is the best alternative to Amazon I have found. They don't provide services like SQS, but they do have scalable storage.
I suggest you change that aggressive, pouty tone of yours. I was addressing a specific point in your comment, not disagreeing with you.
Ok, that doesn't include EBS and bandwidth. So let's throw in 1TB of outbound data (less than the 2TB that Rackspace includes, but still more than I'd ever use in a month) and a 150GB EBS volume with 1000 provision IOPS. Let's also use the 1 year heavy utilization price while we're at it. Now I'm looking $342/month with $1800 upfront, which amortizes to $492/month.
With AWS I also don't need to stick with paying 24x7 for infrastructure. I don't need to keep my CI infrastructure running on nights and weekends for example, and with AWS I have the flexibility to turn that into cost savings.
So I don't see the order of magnitude price difference. So maybe I shouldn't be looking at Rackspace? Hetzner keeps coming up on here, but I don't need a server in Europe. Maybe OVH? Hmm, the comments on Hacker News don't exactly inspire confidence in OVH for something more than a personal server to screw around with. I need to run infrastructure for my business. Do I start hunting for bargain hosting providers on WebHostingTalk?
Ok, back to AWS it is for me.
EDIT: Thanks for the suggestions to look at Softlayer. A server with 1 CPU and 8GB of RAM starts at $399, which is better than Rackspace, but not exactly orders of magnitude cheaper than the c1.xlarge I mentioned above.