- Public clouds
- Internal clouds
- Cloud Services
- Web Infrastructure
- Data Centers
- The Internet
- Near-Term Outlook
- Key Takeaways
- Further Reading
- About Derrick Harris
Two things stand out above all else when looking back on the first quarter of 2011: Infrastructure as a service (IaaS), the epitome of cloud computing, came down to Earth, while big data took flight.
In the case of IaaS, coming back down to Earth isn’t a bad thing; it’s just a sign that the market is settling into a state in which it likely will remain for some time to come. Amazon Web Services continues to lead in terms of customers and innovation — it released Elastic Beanstalk, CloudFormation, Dedicated Instances and an enhanced Virtual Private Cloud service during the quarter — while Rackspace, buoyed by momentum around OpenStack, will be its primary competitor for mainstream customers. As the IaaS market activity illustrates, however, there will always be a case for enterprise-focused providers as well as cloud providers connected to vendors like VMware and Microsoft and pushing highly integrated hybrid clouds.
The big data space, on the other hand, looks a lot like cloud computing did a few years ago: There are so many players and so many terms floating about, it’s difficult for outsiders to get a handle on who’s who and what’s what. For example, where there once was Cloudera offering a distribution of the Apache Hadoop project and some associated products and services, there now are more than a half dozen companies, all utilizing Hadoop in slightly different manners and making slightly different claims. The same goes for NoSQL databases, the proprietors of which still have some educating to do within traditional companies, but whom will have an even more difficult time doing so with the recent proliferation of projects, terms and well-funded startups. None of this is bad news, of course — this type of activity validates the value of the technologies — but it does lead to confusion as lay customers try to determine which “big data” tools are right for their needs.
The other truly noteworthy realization gleaned from the quarter is that Intel won’t be left to die if low-power servers based on x86 processors catch on like the buzz late last year suggests they will. The reason is that Intel is pushing its own line of low-power x86 processors in both Xeon and Atom forms, and even has designed a dual-core server-ready Atom processor for massively parallel startup SeaMicro. If it’s energy usage that’s driving the discussion around ARM processors, Intel sought to prove during the quarter that businesses don’t have to abandon the x86 architecture to achieve that end.
Elsewhere, it was a slow few months in terms of huge announcements or scandalous goings on, but a few big items did set the stage for an exciting year to come. Among those are HP’s impending presences in the cloud computing and big data spaces, whether Juniper can sell a redesigned data-center network architecture and if private clouds built by well-pedigreed, but possibly enterprise-naive startups can actually catch on.