Earlier this year I wrote a comparison of several cloud infrastructure-as-a-service (IaaS) providers, to serve as a starting point for anyone looking to take the plunge into cloud computing. The Infrastructure as a Service (IaaS) market, however, has proven to be one of the most exciting ones in the cloud space, and there have been several important movements, such as changes in pricing strategies, the appearance of new smaller players, and the entrance of some technology heavyweights. All these changes led me to two conclusions: first, the original comparison needed an update, since the relative positions of the main players in the space may have changed; second, it would actually make more sense to split the different IaaS providers into two groups: the “top-of-mind”, most well-known and larger companies in this space, and all the new contenders and upstarts.

Once again, my focus is going to be providers whose services can be bought straight from the web, with a simple credit card, without the need for any interaction with salespeople. The rationale behind this is that if someone is going to run experiments and get into the cloud for the first time, they want to do this without any kind of hassle or friction.

Once again, I tried to look at dimensions that translate the main promises of cloud computing, such as scalability, cost reductions, service levels, and others. But I also tried to focus on the main user concerns: security, ease of migration, true reliability levels, and so on. I have also tried to group the dimensions into related topics, so that comparisons would be easier. So here are the dimensions:

Cloud promises

  • Cost reductions / optimizations
    • Variety of Pricing Plans – The more variety offered (hourly, monthly, etc), the better a provider is considered.
    • Average Monthly Price – Estimated cost in US$ for a 1 CPU, 2GB RAM cloud server (or the nearest best option), averaged over datacenters for companies with location-based pricing, and averaged over Windows/Linux servers. When available, hourly pricing was used, based on 730-hour months. Otherwise, monthly pricing was used.
    • Cost of Outbound Data Transfer – The cost, in US$, for each GB of outbound data sent from the server. Companies that offer a per second (Mbps) connection for free have costs listed as zero.
    • Cost of Inbound Data Transfer – Same as above, but for inbound data.
    • Storage Costs – Average cost per GB of persistent storage. While most VMs will come with some storage included, if you want true persistent storage you have to do it externally.
  • Scalability and Automation
    • Scale Up – If it’s possible to scale up your servers automatically, by adding more disk space, RAM or processing units.
    • Scale Out – If it’s possible to quickly and easily deploy new images based on existing VMs.
    • APIs – If the company offers APIs to interact with the servers or not.
    • Monitoring – A 3-level subjective scale measuring the easy availability of monitoring tools:
  • Poor – Companies that have no monitoring/alert solutions integrated, requiring the deployment of third-party tools or that extra services be purchased
  • Average – Companies with very simple integrated monitoring tools (few indicators or no alerting)
  • Extensive – Companies with very complete integrated monitoring tools offered for no additional cost
  • Choice and Flexibility
    • Number of Datacenter Locations – The number of different datacenter locations where cloud servers can be hosted.
    • Number of Instance Types – The number of different available instance types, in terms of RAM, CPU, disks and so on.
    • Supported Operating Systems – The number of different supported operating systems (regardless of version) available as pre-configured images.

User concerns

  • Security Features
    • Certifications – If the vendor has compliance- and security-related certifications, such as PCI or SAS 70.
    • Protection – If the vendor offers the possibility of protecting servers with firewalls and other security functionality. A 3-level subjective scale:
  • Poor – Companies that only offer the most basic security features (such as a basic firewall), or no features at all
  • Average – Companies that offer a more advanced mix of security features.
  • Extensive – Companies that offer not only several security features, but also some security automation.
  • Ease of Migration
    • Open Standards – If the vendor employs or supports open standards in cloud infrastructure.
    • VM Upload – If the vendor supports uploading your own machine images (made locally) to the cloud
  • Reliability
    • Service Age – How long the service has been around.
    • Service Level Agreement (SLA) – The uptime SLA offered (regardless of past performance), in percentage points.
    • Support – A 3-level subjective scale:
  • Poor – Companies that only offer on-line forums for free; any other support must be paid
  • Average – Companies that offer a single type of 24×7 support for free (either phone-based or on-line chat), in addition to forums
  • Extensive – Companies with multiple support offerings included in the base price

So we have jumped from 14 dimensions to 19, but they can give us a much better picture of the different providers. I’ve also increased the number of companies in the comparison from 11 to 16. This is a result of adding seven new companies (Microsoft, IBM, AT&T, Google, HP, Lunacloud and Tier3) and removing two of the existing ones (GoDaddy, who isn’t positioned as a IaaS provider, and Reliacloud, who removed the information I needed for the comparison from their website).

Here are the two comparison tables in thumbnails that you can click on for a larger view: (Cloud Promises and User Concerns):

Click to see comparison of Cloud Promises.
Click for how they rate on User Concerns.

You can also download the IaaS rating charts as an Excel spreadsheet.

Summary

Since the last time I made the comparison, the first thing that called my attention was that the information available on the websites of IaaS providers has increased greatly. It was much easier to find pricing data, SLAs, and so on than it was before. In fact, some of the providers that I left out of the first comparison, such as IBM, due to a lack of available information, have now been included. I believe this marks an evolution in the market, especially increased competition, which has led to better availability of information.

Once again, there is a wide variation in price ranges, from under US$ 50 to over US$ 200. There was, however, a widespread reduction in prices. This is evidence of two factors: the increased competition in this space, as well as economies of scale that large IaaS providers enjoy and that they can give back to their customers.

A couple more interesting trends are that most providers are now offering both “scale-up” and “scale-out” features, that is, the ability to dynamically increase storage, RAM and CPUs on a single server, as well as the ability to easily clone servers based on predetermined images. This is the natural evolution in this market, since this scalability is one of the greatest cloud promises. Following this trend is the way that many providers are now allowing customers to build entirely customizable cloud servers in terms of resources, opposed to the fixed instance types that Amazon and Rackspace adopt.

Finally, I’m also seeing many providers basing their solutions on VMware technology and allowing customers to upload their own images to the cloud, which can be a very interesting possibility, especially for enterprise customers.

Once again, this comparison is far from authoritative, but it’s meant to serve as a good starting point for anyone trying to easily see the similarities and differences between cloud providers. This can help both the newcomers to the cloud as well as people looking to change their current provider. It was based on information publicly available on the web site of the providers, so there might be variations in contracts for specific customers. Finally, if there is additional information, dimensions or providers that I’ve forgotten along the way, please share in the comments.

Sources