Infrastructure as a service is poised to make the jump from cutting-edge technology to standard enterprise practice, given the latest survey results showing increased deployment in working data environments.
According to Enterprise Strategy Group, at the behest of VMware, more than two-thirds of mid-level enterprise firms are using IaaS for mission-critical workloads. Of that group, more than 80 percent are running production environments on IaaS, as opposed to test and development, while 70 percent are using it for storage. Another significant finding is that 78 percent report using the public cloud for IaaS, but say it is crucial that outside platforms be compatible with internal private cloud and virtualized infrastructure.
All this is putting IaaS on track to become a $4 billion industry by 2015, according to In-Stat. This is likely to propel other service models like platform as a service (PaaS) and software as a service (SaaS) because it affords the opportunity to deploy entire application environments rather than just piecemeal services that then have to be integrated into the broader enterprise architecture. The technology will be particularly beneficial to small businesses, primarily those involved in hospitality, health care and social services, which are expected to provide half of IaaS revenues by 2015.
To gain widespread acceptance, however, the IaaS industry will need to foster a higher degree of interoperability among providers. The Open Data Center Alliance is hoping to move this along with a new set of usage models intended to remove some of the barriers surrounding security and management that currently exist between individual services. The models include guidance for compute infrastructure as a service (CIaaS), which covers basic principles, metrics and architectural frameworks for IaaS delivery, as well as service orchestration and a basic commercial framework to formulate service agreements and other business arrangements.
A key driver in the shift toward IaaS is the need to accumulate the resources necessary to handle Big Data loads. Fortunately, the leading open source IaaS platforms are working feverishly to employ the Apache Hadoop platform that makes it easier to pool the necessary resources to handle large data volumes, reports linux.com's Libby Clark. OpenStack, for example, has added an Open vSwitch patch to the Folsom release to boost Hadoop performance, while CloudStack is turning to dedicated resource pools in order to guarantee Hadoop workloads. Meanwhile, Eucalyptus is aiming for full Hadoop integration to ensure near-native performance in open cloud environments.
IaaS is undoubtedly a powerful asset in the drive to transform infrastructure from a monolithic hardware platform to a dynamic data environment, but that doesn't mean the enterprise needs to rush into IaaS deployments just to keep up with the trends. As resources become more malleable, decisions regarding infrastructure should focus more on solutions rather than mere capabilities.
It's important to keep in mind that most IaaS environments will span both internal and external infrastructure, so it will still require a fair amount of hands-on guidance and oversight from existing IT personnel.
Yes, it will add a great deal of flexibility, but it will also present new challenges in the drive to maintain control over data environments.