Dan's Take

'Intelligent' Analytics for VMware

SIOS iQ takes on the "I/O Blender" effect.

• By Dan Kusnetzky
• 07/28/2015

Jerry Melnick of SIOS Technology Corp. came by to introduce me to the company and to its SIOS iQ machine learning analytics platform for VMware-based computing environments. During the discussion, it came out that the company also now owns a powerful clustering technology.

Virtual computing environments can be extremely complex. There are many elements of technology, living in many layers of software and operating simultaneously; this can make it difficult to find the source of a problem. If you step back to look at the environment, you can see there are a number of OSes, database managers, application frameworks and other software components executing. It's just not easy to know what's going on. SIOS has developed SIOS iQ to address these challenges.

SIOS believes a tool can discover the relationships among and between these elements, then learn how they typically work together to make it possible to quickly and effectively discover anomalies before they turn into problems. This, of course, requires that the tool be able to quickly gather data and analyze it.

Host-Based Caching
Virtual environments that support many different workloads use storage differently than single-function computing environments. The needs of different workloads, such as database, virtual desktops, Microsoft Exchange and SharePoint, and other applications end up creating problems for storage. The industry catchphrase "I/O Blender" has emerged to describe this problem. SIOS iQ includes a tool to analyze storage usage of virtual workloads, and help enterprises make the best use of their solid-state drive storage or in-memory caches.

If an IT administrator is looking at a large group of virtual machines (VMs), it can become quite difficult to know what each of them is doing—or if it's doing anything at all. SIOS iQ has the ability to monitor VMs to determine if they've been idle for a period of time.

I'm hearing a lot about different types of monitoring and management tools these days. It's become clear to just about everyone that host systems running many virtual environments create a very complex environment, one that's tremendously difficult to monitor and manage.

Mastering the Tricks
Many vendors offer monitoring tools for VMware-based environments, marketing those tools as "solutions." But there are several key tricks that must be mastered if the tool is going to live up to its name and really be a "solution."

One trick is gathering the right data, from the right software elements, at the right time, to capture what's happening. This means obtaining the right data from the host system, the hypervisor, the storage system, the networking system and other important operational metrics.

Another trick is having the right tools to perform data analysis. This means being able to sift through operational data quickly and efficiently to learn what's normal and what's not.

When something appears unusual, the next challenge is finding a way to present that information to an IT administrator in a way to make it actionable and useful.

Dan's Take: Complex Environments Need Complex Tools
SIOS has developed machine intelligence algorithms designed to learn from the stream of machine and operational data to create an understanding of what's happening. The data and the understanding generated by the machine intelligence engine is then presented to IT administrators in an easy-to-read way.

SIOS appears to have a useful tool for VMware based environments. But so do many other suppliers of monitoring and management tools. It isn't clear to me who has the smartest machine "brain," and who's offering something good enough to get by. That will come out in time.

If your company is challenged by a complex VMware environment, seeing the SIOS demo and learning more about what iQ can do is in order.