Reflex Shows What's Going On Inside Your VMs

VMware Pricing Controversy: Exclusive User Research

As VMware environments get larger and more complex, they begin to give up some of the cost savings they won through server consolidation--unless a strict management hand is held over them. Reflex Systems, an ambitious 11-year-old company, has put more muscle in that hand by adding vCapacity to its already muscular Virtualization Management Center.

Reflex Systems started out in 2000 as Reflex Security, if that's an indicator of how much times have changed. That's allowed it to take the unusual path of claiming that it can manage both virtual intrusion protection and managing virtual machines, "no mean feat," according to Rachel Chalmers at market researcher the 451 Group in her March 2010 report on the firm.

The automated provisioning and monitoring of virtual machines is a well established field, with both VMware and the third-party suppliers who compete in its environment, such as Veeam, DynamicOps, and the Vizioncore unit of Quest Software, active in the field. Reflex, however, may be a step ahead as it adds vCapacity to its own management center.

VMware itself only recently brought its first capacity management capability to its environment with vCenter Operations. Virtual machines are historically configured liberally, with extra virtual memory and storage to avoid running up against too narrow a set of limits during intense activity periods. The fact that they seldom drew down available physical resources meant that over-provisioning in the virtual world was a less costly guarantee of operations than in the real world.

Having encouraged the practice, VMware got itself in trouble when it proposed recently that future pricing for use of vSphere 5 be based on the configuration of virtual memory in a virtual machines, and whether that memory exceeded limits that it had set for VMs under different editions of vSphere. It was a calculation that many customers saw as a source of price inflation.

In fact, VMware may also have been implicitly lobbying for greater use of capacity management tools that would allow customers to come closer to right sizing virtual memory to virtual machines based on use, then managing them based on trend lines. At least, it claimed its future pricing strategy would be based more on use of virtual resources and less on physical assets, such as CPUs.

Thus, Reflex made a timely entrance Monday onto a stage currently in controversy's spotlight: virtual capacity management. Despite good intentions, it has thus far remained one of the darker arts of IT science.

Capacity management for VMs is one way to move away from old data center over-provisioning of physical resources to "leveraging virtualization to enable the design of next generation architectures, such as private clouds," said Reflex CEO Preston Futrell in an interview. With vCapacity, virtual machine manager can both provision and manage, monitor current needs, and plan for future additions to the virtual infrastructure.

Reflex already has a product suite that offers security monitoring (vTrust), virtual machine monitoring (vWatch), and configuration (vProfile), so vCapacity potentially fits right in. The addition of vCapacity, however, brings a combined view of each tool's data shown in a browser-based dashboard, something that wasn't available before. Previously the tools presented individual views through an on-premises Windows client.

The new combined view, as shown by Futrell and Mike Wronski, VP of products, in a demonstration, is offered in a browser-based dashboard on the Web. The dashboard shows the number of virtual machines running, number of hosts, number of CPUs utilized along with their number of cores, the amount of storage used, the density of virtual machines on hosts, and recent peak usage of each resource by server cluster.

VMware Pricing Controversy: Exclusive User Research

Other detailed views can be summoned, moving from a general, multiple VMware vCenters view down into the operation of individual virtual machines, a reporting technique that enables troubleshooting on the virtual environment when a host shows signs of slowing down, Wronski said in an interview.

VCapacity collects data on operations and establishes trendlines with it, giving virtual machine managers a tool with which to project when they will run out of capacity and steps they can take to avoid doing so. The dashboard can be configured to display information that the customer is most interested in and can be tuned to send alerts upon reaching certain thresholds, Wronski said.

Complex event processing, where the system watches a stream of software events and concludes some mean is being deviated from, gives vCapacity analytical capabilities after it's collected enough historical data to establish trends. "Trends start to show up in days, not weeks," Futrell said.

The dashboard approach is meant to give business managers, the owners of virtualized applications, the ability to see key performance information, as well as give IT managers key operational information. Views can be customized on the dashboard for each group, Futrell said.

The largest environment in which Reflex has tested vCapacity and the Virtual Management Center suite is 2,000 VMware ESX Server hosts. The average size of a Reflex customer today is 300-400 hosts, he said.

VMware's vCenter Operations is limited to 500 hosts; Reflex can move beyond that figure, said Fuller. But Reflex's main competition in capacity management is not VMware but the homemade tools that most IT shops are using in trying to perform the task themselves, he said.

VCapacity relies on a configuration management database as a single source of data on virtual machines and VMware's vCenter Operations does not rely on one. Customers will have to decide for themselves whether that's an important differentiator.

For more on the vSphere 5 pricing controversy, see VMware Blinks On RAM-based Licensing Costs and VMware Pricing Controversy: Exclusive User Research.

InformationWeek Analytics has published a report on backing up VM disk files and building a resilient infrastructure that can tolerate hardware and software failures. After all, what's the point of constructing a virtualized infrastructure without a plan to keep systems up and running in case of a glitch--or outright disaster? Download the report now. (Free registration required.)