@CloudPhysics https://www.cloudphysics.com/ , who presented to us at the #TFD11 created a unique idea geared toward the idea that the general health of your virtual infrastructure is a complex and multi-layered thing. vCenter, though better than it has been in the past, still gives quite limited insight into the true health of your infrastructure. It’s a question of granularity. VMware collects tons of data, often if you’re command line savvy, you can dig into those stats, and see how things are behaving. But, it’s so difficult that they’ve really made less than ample effort to expose these health related statistics.
Enter an ecosystem of players in the field who go at this difficulty each with differing approaches. I could list off a few, some with monetary aspects toward presenting a cash based value to the environment, thereby making the cost limiting or cash lean factors, valuable to the management of the virtual infrastructure. I think that was a compelling argument, though it can get a bit overplayed.
Other approaches rely on on-prem virtual machines that act as collectors. They cause effect into the environment by simply existing. In a big way, it’s like the Observer Effect, (link) which is a scientific term that states the act of observing changes the task being observed. If you’re using an electron microscope to view the effects of an atom by some criteria, the observation of the atom by the electron microscope will quite literally change the behavior of that atom. This easily can be exposed by a VM which grows as it collects and watches the environment, effecting the storage of data, the processing of the analytics against the overall memory and processor used, the network utilization, and the overhead against vCenter.
I do believe this to be significant in its value toward the rest of the pack.
Enter @CloudPhysics. They’ve created a manner of exporting only performance data from vCenter to an Amazon EC2 centralized repository. This is light data movement, and as such, the effect on the virtual infrastructure is quite minimal. One of the keys to this platform is that the data being collected is only performance data, and none of the business critical data which would contain sensitive information that relates to the business itself.
Then, the #CloudPhysics app takes that data and delivers it in a dashboard that easily presents those analytics into a well-constructed web interface. Anyone with rights into that interface can see whatever health information is relevant to them.
Even more interesting to me was that this data can be easily cross-referenced against the collected data of all the other companies who are contributing these analytics toward the aggregate. How does my health compare to those of the environments similar to mine? And again, because they’ve constructed an environment of corporate sensitive data never being collected, an organization can feel secure in uploading and comparing this data to peers.
To be clear, today, @CloudPhysics is only built to collect vSphere data into their environment, but plans are under-way to create the capacities to collect that data from other virtual environments. This roadmap will not only allow you to collect but to compare your environment to other vSphere infrastructures, but as well, to potentially Xen or openstack environments. Where best to sit that VM? CloudPhysics can help.