Toronto-based server workload migrator and monitoring vendor PlateSpin jointly hosted a Webinar Wednesday with its Lowell, Massachusetts-based virtualization partner Virtual Iron Software, and shared several of its strategies, which include keeping a stern eye on your data centre’s resources and ensuring the easy portability of different server systems.
IT World's Video Library
John Stetic, senior director of product management for PlateSpin, discussed its workload management solutions, which fall under its two software offerings, PowerRecon and PowerConvert.
The former assembles a workload profile that gauges which resources are used, special applications present, the host name and owner, and the costs involved, said Stetic. “It collates data in a non-invasive way that can be used for analysis,” he said. “It offers visibility of your data centre, of the workloads and resources available, and how well-matched the workloads are to the type of server.”
Workload portability—ensuring that workloads are allocated correctly for optimal data centre consolidation, performance, and cost efficiency—comes into play with the latter offering, PowerConvert. It offers portability between different types of servers, be it physical servers, blades, or virtual environments. “You can also capture an image of the workload and send it out to any environment,” said Stetic.
These solutions work well with one of its virtualization partners, Virtual Iron Software. (PlateSpin’s technology also integrates with products from other vendors, such as the customers of virtualization giant, VMWare.)
Chris Barclay, Virtual Iron Software’s director of product management, labels the company an enterprise server virtualization and management vendor. Its product, which is based on the popular Xen hypervisor, aids businesses in managing their multi-server environments. “There’s often physical and virtual servers in the same environment. You need to balance out your resources and workloads in a diverse environment,” said Barclay.
Without undertaking such a balancing, he said, data centres run the risk of mangling their potential. Said Barclay: “With a large resource supply, if you underutilize it, there’s wastage. But if it’s overutilized, there’s risk.”
Virtual Iron offers an easy-to-use Web-based interface, and live provisioning, which allows the data centre administrator to work without having to take the server offline, according to Barclay. An enterprise-class management program offers virtual service management capabilities.
CASE STUDY: Memory Sizing for Server Virtualization.
If you have to update the hardware, cloning the images of the workloads out into a virtual snapshot allows the admin to work without taking it offline, which is especially helpful during patch management, said Barclay.
These live images can also be used to move content around onto different hardware. “You can basically restart your virtual machine in any environment,” said Stetic. This comes in handy if one’s hardware goes down. And, if it did, the PlateSpin products have built-in data recovery properties to retrieve the information.
PlateSpin’s solutions are generally implemented by value-added resellers (VARs), who install (or advise the customer to implement) PlateSpin’s set-up for optimal server performance. The company has seen success, said Stetic, due to the popularity of mixed server environments, and the increasing demand for portability between them.
PlateSpin’s typical process, executed by the VARs or the customers, involves inventorying the data centre holdings and collecting it centrally. This is an especially important step for not only those who are new to virtualization (and need to know what they’re getting into, and where in the data centre would be the best fit), but for long-term care of one’s server holdings, which can be monitored by PowerRecon.
Said Stetic: “When switching to a shared resource environment, you need to see what the cost allocations are, and application visibility. It’s easy to get new workloads online, and that can create virtual machine sprawl, so it’s important to know what the costs are (and whether the workloads are allocated accordingly).” The program will also provide chargeback and reporting capability to aid in this.
Once the servers and their output have been well-documented, the implementation begins, which is eased by a user-friendly install (and, if the customer doesn’t find it a good fit, an easy uninstall, too) with easy transfers of workloads among machines. “You have to right-size your resource allocation for optimal configuration,” said Stetic.