Internap Network Services Corporation (NASDAQ: INAP) this week announced a dual-cloud compute platform that leverages both open source and proprietary cloud technologies.
Internap will be leveraging the open source OpenStack platform and VMware as part of its public cloud offering. By providing both stacks, Internap is aiming to provide more choice and data resiliency for cloud customers.
"CIOs' reception to the dual-hypervisor stack is positive, most have virtualized internally using VMware as a server consolidation play, and they are looking at a cloud that supports VMware as well as open source," Paul Carmody SVP, product management at Internap told InternetNews.com. "We know that just like Linux grew to replace Solaris in the enterprise, people will get used to open source alternatives, and enterprises want to choose a vendor that will allow them to migrate in that direction."
Carmody noted that the CIOs he has spoken with are not necessarily familiar with OpenStack as a brand, but they are interested in open source hypervisors in general. OpenStack started as a joint effort of NASA and Rackspace in 2010, and it now has more than 40 contributing member companies. The OpenStack Cactus release came out in April, providing new authentication and API features.
While Internap will provide both VMware (NYSE: VMW) and OpenStack public cloud options, Carmody said it is possible to migrate from one to the other. He added that from a backup and failover perspective, enterprises can do a number of things to ensure data resiliency.
"Enterprise folks have gotten comfortable with the idea the VMware has certain high-availability characteristics that allows for dealing with node outages," Carmody said.
Carmody noted that features like vMotion, which enable virtual machines to move from a failing node, are part of the solution. He added that it is also possible to set up VMware in a high-availability mode.
"With OpenStack and Internap, we're focused on availability and support," Carmody said. "A lot of the cloud failures that have happened recently in the marketplace were related to network issues as opposed to hypervisor issues"
Amazon in April suffered though several days of cloud service disruption. The Amazon outage was triggered by an issue in a northern Virginia data center.
"Internap starts with the network up, our heritage starts with the network, and we also offer managed networking services, including managed routers," Carmody said.
One of the other things Internap is providing for both OpenStack and VMware cloud environments is content acceleration.
"We poll for the Internet for availability, congestion and latency," Carmody said. "We developed routing tables based on that information that we inject into our routers that guarantee the optimal route."
Internap also has its own TCP acceleration technology, called Zip. Carmody explained that TCP determines the window size for how much information you are sending in a particular communication, from the source to the end user. TCP assumes that the connection is bad, and over time it increases the window size.
"We have the ability to accelerate the traffic by doing larger tcp window sizes at the outset which results in better throughput," Carmody said.