Blog Post

6 IaaS providers you should know (but may not)

When people think Infrastructure as a Service their first thought is Amazon(S amzn) Web Services. That’s fair given its early mover advantage and size. But AWS is not the only IaaS in town.  The thing is some much smaller cloud providers are not trying to compete with AWS on sheer size and breadth of services but more on specialized features and services or by focusing on a smaller market. The feeling among cloud analysts is that a smaller provider can out-innovate the big guy provided they focus that innovation and don’t try to be all things to all people.

“While AWS, Rackspace(s rax), and Microsoft(s msft) are on everyone’s short list, you have look at every IaaS player to determine fit.   The objective is to move from the business, to the requirements, to the technology.   If you do that, you need to be a bit more open-minded when it comes to selecting a cloud provider,” said David Linthicum, Gigaom Research analyst.

Given that assessment, here’s list of a few IaaS offerings that have cropped up often in my conversations over the past few months.  You may not know them by name but they may be worth a look, depending on your needs.

1: Tier 3. This small cloud provider got outsized kudos from Gartner’s rankings for what Gartner(s IT) cloud analyst Lydia Leong described as a “good balance of manageability and control for an IT audience and productivity-oriented stuff for developers.” For developers, its Iron Foundry PaaS — basically a version of Cloud Foundry for the .NET world — is a strong entry and Tier 3 supports popular devops tools including New Relic, Pingdom, Logstash, and Octopus Deploy.

2: ProfitBricks.  This company zigged where Amazon zagged, offering scale-up architecture — you can rent gigantic instances of compute power — up to 62 CPU cores and 240 GB of RAM — which suits traditional big database applications as well as some NoSQL scenarios. “What I like about them is instead of imitating Amazon EC2 like most of the newbies, they leapfrogged AWS with their scale-up architecture, dynamic instance types, InfiniBand and SSD based storage,” said Gigaom Research analyst Janakiram MSV.

3: Virtustream. For enterprise accounts needing cloud infrastructure to run enterprise SAP applications  Virtustream is probably a great option. So great, that SAP invested in Virtustream. The company was smart to pick an application it excels at and then target the heck out of that audience. Virtustream has 100 production large-scale SAP implementations running –that’s impressive,” said Forrester Research analyst Dave Bartoletti.

4: DigitalOcean. For newbies who may not know what an instance is but want to build stuff on rented modern infrastructure. DigitalOcean is “wicked cheap, adding customers fast and doesn’t have a ton of legacy infrastructure so they’re the first time of where IaaS prices are headed,” said Carl Brooks, analyst at 451 Research.

5: Internap: Has a nice mix of cloud and bare metal capabilities. And its recent purchase of iWeb may make it home to a lot more small and medium-sized companies. Brooks said Internap(s inap) is “turning colocation into cloud long-term, and does interesting stuff around DCIM [Data Center Infrastructure Management].” Customers like the mix of bare metal and virtualized options. Data analytics company Exelate uses Internap a lot because it’s able to define its own server specs, network topology and storage arrays, said Exelate co-founder and CTO Elad Efraim. 

6: GoGrid: Offers very enterprise-y service level agreements (SLAs) that distinguish it from the pack, said Forrester Research(s forr) analyst Dave Bartoletti. And with its GoGrid Exchange ISVs can license and package up their applications to run on GoGrid, making deployment easier for end users.

This is hardly a scientific sample. Please use comments below to propose other, smaller IaaS providers and specifically why they deserve a look. Remember, AWS, Rackspace, Microsoft, HP need not apply for our purposes here.

Feature art courtesy of Shutterstock user Matthias Pahl

33 Responses to “6 IaaS providers you should know (but may not)”

  1. 2 of these names (GoGrid and Internap) can be accessed through a single sign on into ComputeNext Marketplace with 24 different machines for GoGrid and 84 different machines for Internap.

    They have different platforms, but are unified through a sigle API. Joyent mentioned in a comment is also available with 11 machines.

    In terms of geographical location Internap offers services in the Netherlands while GoGrid is US based as well as Joyent.

    There are currently a total of 23 public IaaS providers across 23 countries.

  2. Hi All,

    Check out Peak, I work for them so drinking my own koolaid here – but outside of customer service with an enterprise focus we do in essence hosted private cloud. Public is not our focus. We also have a patent pending for L2 that allows you to bring your on prem IP schema into our fabric with no integration headaches. We shoulda been on this list Barb! :)

  3. Barb. IT is good to see an ever growing list of organisations getting into the IaaS list and providing services, but I think you need to add some other key criteria that I think a lot of organisations overlook when making a selection.
    e.g. – How fast can the client scale up or scale down the services. Can it be done by the client.
    – Where are the services being provided (country, state)
    – What is the performance to this site (response time)
    – Is the DC backup up and has automatic failover
    – Is the client’s data backed up, and is it available at the DR site
    – What security services are included in the IaaS, or can they be added (e.g. Firewall, IDS, etc)
    – What performance management is included and accessible by the client, or can be added
    – Is there any acceleration tools available on the connections
    These are some of the areas I feel are needed for clients to make a call on which provider to choose, and not just on who is the cheapest, and there are probably going to be more. But, every client will have a different weighting for making their choice.

  4. Wondering why Softlayer, an IBM Company is not even mentioned despite the publicity around the $2B purchase in July, 2013, 22,000 customers and growing, and the recently launched advertisement to inform the public (support 270,000 more website than the nearest competitor, host 30 percent more top website than any other hosting provider in the world, server 24 of the top 25 Fortune 500 companies).

  5. Not an IAAS provider per se, but Ravello Systems is a service that allows the importing of applications/VMs from data centers as is, and deploying them on multiple public clouds.

  6. It’s great that there are many cloud services out there, but how many of them have ‘back door’ access by the National Security Agency, or extra ports in their routers for NSA to take copies of all of their traffic ? Is any data stored on cloud providers in the USA, Britain or Canada safe from examination by their governments ?

  7. Nice list. We mainly use Digital Ocean and AWS. But I could think of a few more that are rising. Nephoscale, Linode and Softlayer come to mind. Where things are heading now, multicloud is the way to go. With New Relic, or it’s pretty straightforward to manage servers on multiple clouds these days.

      • Take a look at Digital Ocean and NephoScale! We use NephoScale for their top tier Private Cloud and Hybrid Cloud solutions and we use Digital Ocean for testing beta applications. Both providers have great technology and awesome customer support.

  8. How do you rate Vmware as an IaaS player? Most of the enterprises have chosen vmware virtualization and the cloud or IaaS is built on the virtualization platform. How come vmware or vmware based service providers are not making the cut? what is missing? I am not sure if people can build infrastructure without server virtualization. Are they building cloud on top of vmware?

  9. greg bears

    good stuff.

    digital ocean is doing very well almost amazingly so.

    internap haven’t done anything with the (cool) Voxel technology they bought which is disappointing.

    joyent is pretty cool and they have a very engineering-heavy background. i am surprised you didn’t mention them.

    another one that i think you have kind of forgotten about would be logic works. they certainly deserve a mention no?


    • Thanks for the mention, Greg.

      I’m the most biased person in the world to agree that Voxel has some cool stuff :)

      Stay tuned though — we have done a lot of stuff with it post-Voxel in Internap-land, and are going to make some interesting announcements imminently.


      Raj Dutt
      SVP Technology/Internap

  10. AWS has provided so many options that sometimes it confuses the buyer to settle down with an optimum and cost effective package. the IaaS providers metioned by you are worth awhile, though it still needs a finall analysis of requirement and sometimes training the employees to handle the new IaaS that we bring in.
    Concerning all these issues, we at Fux7 Labs have been creating solutions. Fo startups, we also have free doubt sessions here

  11. Barb: It’s good to share these (I have dabbled with a few but may I know have not heard of them). Are there any others on your list to share? I’d be curious to read anything you had on customers experiences (or analysts opinions) on Joyent

    • ansgartrollgamesde

      I’m a fan of (german) dedicated servers and I’m using them since 10+ years for our browsergames (,, I never understood why someone would pay a crapton of money to use unstable and unflexible AWS VMs. Because of AWS I hated virtualization and IaaS.
      Then, 1.5 years ago, I discovered Profitbricks. The description in the article doesn’t really describes why Profitbricks is awesome: you can build a server like lego. You want a Server with 1 CPU, 80GB RAM and 5TByte storage? No Problem. You need 400 x 1 core, 250MB RAM mini-VMs connected via a local network? Just draw it together on their website and click “start”.
      And then there is this: if one of your VMs needs more CPU or RAM, just ADD IT WHILE THE MACHINE IS RUNNING! There is no downtime, and this makes AWS and all the maintenance stuff you have to build to get around the horizontal scaling problems, look outdated and dumb.
      There are downsides at Profitbricks, but not many. The only thing that comes to my mind is that it is more expensive than running your own dedicated machines (but still much cheaper than AWS) and sometimes their deployment process gets stuck and you have to call the support. But we switched nevertheless and besides the flexibility the argument that got us was the support. You can call them 24/7 and they always understand exactly what you are talking about. Say “I got ‘microcode: failed to load file amd-ucode/microcode_amd_fam15h.bin’ in my dmesg, WTF?!?” and they will call you back in 10 minutes to tell you.

      The other provider on the list I used is digitalocean because they advertised with “fast SSD Storage”. It is really fast. But the killer feature for me is that you start your VM in 55 seconds. When my boss yells at me that he needs his new, completely untested app online ASAP, I don’t want this on my servers. So I start a so called “droplet” and 5 minutes later I have a running LEMP stack where I can throw all the stuff at, that I’d hate to see at my precious other servers. I use digitalocean as a throwaway server provider ;)
      The main reason why we’re not using digitalocean as our main provider is that Profitbricks is more of a complete solution to us. I think that digitalocean has the faster storage compared to Profitbricks, which is an important argument when running SQL databases. At Profitbricks you get a built in firewall on your “lego-network interfaces”.