The larger your site gets the less you're going to need to rely on "hourly" scaling (as your traffic models also become more predictable). There's also the question of uptime, which we've all seen is not a guarantee from just using Amazon.
The larger your site gets the less you're going to need to rely on "hourly" scaling (as your traffic models also become more predictable). There's also the question of uptime, which we've all seen is not a guarantee from just using Amazon.
Can you qualify this? The cost is lower than Amazon and Rackspace Cloud.
There are dime-store VPS providers out there that are technically less expensive, but they're usually not able to provide the same level of service. Linode always seems to provide exceptional customer support.
Some VPS providers specialize and have a good history of both uptime and support, and provide it for less than the big-name competitors. I've found many to be exceptionally reliable with good support, and there are communities where you can find reviews and status on such providers to back up their claims.
And in the end, realistically, if you're doing small scaling by the hour you probably don't need the biggest providers in the game. A small service provider for a small company is often a great fit.
You can check out reviews on http://www.lowendbox.com/ and http://www.webhostingtalk.com/ and compare prices, services and reputation. In general, a standard small-sized VPS will be half the cost of Linode, meaning you can buy double the VPS's (on different providers, for example) and get double the capacity and redundancy for the same cost. If you spread them out by region you can implement cheap software GeoIP load-balancing and serve the content by servers closest to your users.
This comes from my own experience running my own small scale VPS host for a few years, with a purposely small customer base (mostly people I knew personally).
Look at WHT and follow it for a while. That market is crowded, a race to the bottom in pricing AND support, and is full of drama. Some examples I've seen:
* Lack of experience in production systems with no backup procedure... some systems go down, they loose all data, entire business caves.
* Super tight margins and low capital... fall behind on a build, their provider shuts them down (most these guys are leasing dedicated servers or reselling from another provider)
* Don't know who really is in control... reselling from someone and don't know who? On dedicated servers with a provider who is known for a problematic network?
* They're some 16 year old and their dad grounds them from the computer so they can't answer support tickets.
* Don't understand the technology they're using for virtualization and get noisy neighbors, don't properly secure exploits, don't understand the inner working of their COTS control panel, etc.
* Look up the history of Lxadmin. VERY popular control panel package, company behind it, but most coding/knowledge behind the CEO. Big exploit found, 100k servers exploited, CEO commits suicide. Company is dangling. Believe licensing servers went offline. All providers using them were screwed. They were using COTS software, didn't know their own stack, left scrambling.
Stick with providers that are a real company. Employees of their own, support staff of their own, own their own equipment, etc.
Linode is great, EC2 is great, and there are several other providers.
Run away.
What are some of the others?
Rackspace, HP Cloud, VPS.NET, Gandi, RIMU, Slicehost (back in the day), Webbynode, Liquidweb, StormOnDemand (though a part of Liquidweb), Burst.net, PowerVPS, FDC Servers, ...
List could go on. I usually look for whether they have an established history, a site that doesn't look like a crappy overused template, a real physical address, WHOIS info on the domain, Google them and research, whether they have a real status site, maybe lookup if they have an ARIN number, traceroute to their site, etc.
The last two.
OpenVZ is more OS virtualization, there is less isolation between instances.
OpenVZ usually has "guaranteed CPU/RAM" and "burst CPU/RAM", while KVM always gives you what its assigned. The guaranteed/burst is typically BS. More often these budget providers like OpenVZ because you get greater density, which means more noisy neighbor and taxed resources like disk IO.
However it may be tricky to pick a good host. It is usually a good idea to look for past experiences before signing up. The community is active and helps make that decision: http://www.lowendtalk.com/discussion/5136/best-leb-host-q3-2...
As a hosted continuous deployment service releasing changes and scaling automatically is a necessity.
I think the key here is that this blog is talking about a hosted service which responds on demand to their client's needs on an hourly basis. EC2 has a lot of infrastructure which does this for them (as long as they can automate it) and is specifically built for this, whereas they'd have to write it themselves by interfacing with a VPS API otherwise. So EC2 might be worth the money to them. VPS providers are mostly more focussed on giving you 1 or 2 boxes to use, and then you leave your site running with occasional adjustments in capacity etc, not on spinning up and down servers on demand, which is a bit more specialised, and might require tens or even hundreds of servers.
For a normal site that's not going to be relevant, and a VPS is cheaper, for this site, it seems EC2 is a better fit...
I'll bet you that if they took the historical metrics of their services and their load use and averaged it, buying one or two boxes would easily break even and at the same time not be dependent on a specialized provider or functionality like an auto-scaling VPS.
But that's just a guess.
Working with Historical data means you either totally oversubscribe certain days, or totally underdo it. It has to be automated due to the latest load.
For example on Weekends we typically see 20%-30% the traffic that we see during the week, simply because people are not at work. Scaling up and down is an absolute necessity there.