5 Essential Elements in Building an Agile Data Center
Network virtualization has emerged as a key tool for injecting flexibility into enterprise networks. Nowhere is that flexibility more needed than when an enterprise IT organization moves to embrace emerging technologies such as the OpenStack management platform and containers such as Docker.
Midokura today unveiled a Midokura Enterprise MidoNet (MEM) upgrade of its network virtualization overlay that adds support for both the Kubernetes and Docker Swarm container orchestration engines as well as providing the ability to simultaneously support multiple instances of OpenStack.
In the context of OpenStack, MEM is a plug-in to the Neutron networking module in OpenStack that provides a more robust approach to network virtualization in those environments. All too often, organizations now find themselves running multiple versions of an OpenStack framework that gets regularly updated every six months.
Susan Wu, director of technical marketing for Midokura, says MEM 5.0 now provides a router peering capability that makes it easier for IT organizations to absorb change at the networking level when, for example, they wind up running multiple releases of OpenStack at the same time.
In addition, Midokura is now making it easier to correlate events taking place within MEM and the actual physical networks it runs over. A MEM Insights feature in MidoNet Manager gives IT managers the ability to compare and contrast current and historical events happening across the network virtualization overlay and the network underlay.
“Network managers lose visibility into the underlay when network virtualization gets deployed,” says Wu. “They told us this was an important capability for them to have.”
Emerging technologies such as OpenStack and Docker that seek to fundamentally change the way enterprise IT operates are often a major networking challenge. Legacy applications that depend on network services are not going to disappear just because the organization has embraced another enterprise platform. At the same time, technologies such as OpenStack and Docker containers are in an almost constant state of flux.
In general, adoption of network virtualization and software-defined networks has been slow but steady. IT organizations are clearly interested in the flexibility they provide. But everything from simple organizational inertia and lack of available budget to having to enhance the networking skills required by the IT staff to deploy and manage them means most organizations are shifting to these platforms at significantly different rates.
Of course, shifting to platforms such as OpenStack and Docker almost by definition forces the network virtualization issue. In fact, Wu notes that microservices based on containers, for example, will spawn more network connections than ever. The only way to deal with that increased velocity, says Wu, will be to deploy a network overlay that provides a layer of abstraction needed to logically manage large numbers of network connections being made by those microservices. In that context, microservices such as Docker, along with platforms such as OpenStack, make the shift to network virtualization all but inevitable.
