F5 Friday: Playing in the Infrastructure Orchestra(tion)

I’m sure you’ve noticed that there have been quite a few posts on the topic of automation, orchestration, and infrastructure 2.0. Aside from the fact that an integrated, collaborative infrastructure is necessary to achieve many of the operational efficiencies associated with cloud computing and highly virtualized data centers, it’s also a fascinating topic from the perspective of understanding how network and infrastructure providers are dealing with some of the same issues that enterprise software has long had to face while navigating the enterprise application integration (EAI) landscape.

One of the ways in which vendors like F5 are addressing the need for automation and ultimately the orchestration required to implement a fluid, dynamic infrastructure is through strategic partnerships. These partnerships allow for tighter integration of solutions like BIG-IP (and its myriad feature and product modules) with the emerging infrastructure management solutions coming from industry leaders like HP, Microsoft, and VMware.

A fully integrated, automated IT operations center isn’t, as anyone who’s been through the EAI nightmare, something that happens overnight. You have to start with the automation and management of key components and get them to the point that the automation can be trusted by customers before you expand outward. One of the first components of a dynamic, scalable infrastructure such as those required by cloud computing to be automated through such integration should be the load balancing solution because without load balancing, well, you really can’t implement elastic scalability.

While auto-scaling looks easy from an administrative point of view, what’s really happening under the hood is a more complex set of operations that are initiated when a new application (usually in a virtual machine these days) is launched. The Load balancer has to be notified that a new node (application instance) is available and it has to be inserted into the appropriate pool (farm, cluster) so that the system can begin sending requests to it. That happens through iControl, F5’s standards-based control plane API. That same APi is available for any F5 customer, by the way, and it works the same for the Virtual Edition as it does the hardware versions. Anyone can use the API to develop whatever kind of interesting out-of-band management or monitoring or control application they’d like. That’s always been the case, since the first API call saw the light of day many years ago.  By providing out-of-the-box integration between F5 solutions and three of the most widely used orchestration tools the industry (HP Operations Orchestrator, VMware vCenter Orchestrator, and Microsoft Virtual Machine Manager) a more complete solution for automated infrastructure adaptation with VM provisioning and de-provisioning can be realized.

We’re not at the point where it’ll make your dinner and do the laundry but it’s a decent step forward toward automating enough of the operational tasks associated with a dynamic data center that at least you’ll have more time to make dinner or do the laundry.