November 26, 2008
You could call it Marvin Gaye Impulse, the need to know “what’s goin’ on.” At the IT command center that runs world-famous Hershey Park and other units of Hershey Entertainment and Resorts, they call it being “in the know.”
“We want to do what we can to make the datacenter team in the know,” says Jon Blomeier, datacenter analyst for the Pennsylvania-based company.
Blomeier’s broad purview includes “everything virtual,” and when the team was given the directive of becoming more “in the know” about the virtualized infrastructure, he needed to find some new tools. They’re running VMware ESX on IBM System x3650 servers and were using VMware’s VirtualCenter for management, but “VirtualCenter doesn’t let us look at all the things we want to see,” he says. “We wanted to monitor proactively, and be able to find out why things were happening. We needed something that was gathering data on critical processes like CPU and I/O and memory and reporting in a way that would allow us to react before there’s bad impact on the environment.”
After checking out a few options, Blomeier went with a pair of virtual appliances from VKernel, whose founder, Alex Bakman, says virtual appliances “will prove to be the best approach to simplifying systems management in the datacenter.”
The Hershey team adopted VKernel’s Capacity Analyzer to monitor shared CPU, memory, network, storage and disk I/O utilization trends across its VMware environment, and Chargeback, which provides details about who is using which resources and the costs. VKernel is strictly for VMware at this point, but HyperV is next, Bakman says.
“We’re for the user who needs visibility into their deployment as far as capacity is concerned,” Bakman explains. “How much capacity is left? Where is capacity unused or wasted? We let them go in and see where they can add virtual machines, identify areas of overallocation, see who’s using what resources, and then use that information to make intelligent decisions about how capacity can be better used.”
One of the most useful things about Capacity Analyzer is “it can alert us to problems before they can happen,” Blomeier says. VKernel refers to the appliance’s “predictive analytics” capabilities, serving up data that can be used to adjust resources for maximum performance. Blomeier checks the reports from Capacity Analyzer everyday to get his intelligence briefing on how things are running and where there might be fires smoldering. “I’m constantly looking to see how many days to warning levels on things like CPU and memory resources,” he says. “VKernel makes this information really easy to have.”
“One thing we’re doing with Capacity Analyzer is a summary report that gets e-mailed to our staff every day, a snapshot that shows what’s happening with our entire environment,” illustrates Blomeier. “We can see the current top five capacity bottlenecks, and the top five future ones. It can basically tell us if we don’t do something in 10 days, we’ll run into an issue. We can see all the constrained resources, how many more virtual machines we can fit, the top three data stores that are running out of disk, and the top resource consumer in each category. You can drill down further, to see things like which CPUs are most available.”
Having a firm handle on utilization is a big deal. As Bakman points out, that information can help customers allocate resources judiciously to avoid performance degradation and downtime, avoid buying infrastructure that’s not really needed, and safely increase the number of VMs. “Capacity is very expensive,” he says. “One myth about virtualization is it’s always cheaper, which it is in the long run, but in the short term it’s not. If you can reclaim 20 percent of your capacity -- even if you’re talking about a $100,000 investment -- that is real money.”
Beside the deep visibility and alerting capabilities, Blomeier says ease of deployment and affordability also sold him on the VKernel appliances.
Bakman says simplicity is part of what distinguishes his software’s approach from traditional system management vendors like IBM, HP and CA. “You can download and see the value of our software in about an hour. Because we support the Open Virtual Machine format (OVF), the end user can literally deploy our appliance by going into Virtual Center, giving it a URL, and clicking deploy instantly. A software appliance should have all the simplicity of a consumer appliance. Also, we don’t do like Office and give you 150 features when you only need five. Every appliance has only what you need and nothing more to complicate things. The idea of system management is to aid system management.”
Although Hershey hasn’t yet officially instituted a chargeback program, it is able to use VKernel’s Chargeback appliance to see how different departments are consuming resources.
“The issue with virtualization and chargeback is that the traditional model for expensing hardware does not work,” Bakman says. “All resources are shared, things are shared between virtual machines. Organizations want to know how they charge end users so that they’re able to repay the money they’re spending on hardware. Our idea is to help them meter resource utilization by various corporate centers so they can allocate costs.”
Now You See It …
Andi Mann, who covers virtualization and systems management research director at Enterprise Management Associates, says the biggest problem VKernel solves is the lack of visibility into who is using virtual resources, and how much they are using. “Virtualization adds several layers of complexity, so with normal management tools a lot of granular data about resource utilization gets lost,” he explains. “As a result, physical servers can get overloaded with virtual machines, causing performance to suffer, or physical servers can sit underutilized, not running as many VMs are they can, so consolidation projects are less effective, potential cost savings go unrealized, and costs may even rise as hardware purchases continue to increase.”
VKernel can “drill down into the hidden layers of complexity, showing exactly what resources are being used, how much capacity is available, and what workloads can be co-located. So all resource utilization in the virtual environment is optimized -- stacking as many VMs as possible onto each physical server, but doing so in a way that ensures performance and availability are not affected,” Mann says.
Data from VKernel appliances also can be used to solve some, though not all, of the political issues surrounding shared resources. “With tools like VKernel, IT can point to hard metrics that show exactly which applications have been using which virtual resources, eliminating that finger-pointing over who is using what in the virtual environment,” he says.
Ready for Prime Time?
Still, Mann doesn’t appear convinced that virtual appliances are yet the ultimate answer to the complexities of virtual systems management – an area Mann calls “the major challenge for virtual infrastructures today.”
“I do not think virtual appliances are ready for prime time,” he says, “and in EMA research, virtual appliances were the least desirable form factor of any for datacenter management tools (selected as a preference by only 12 percent of 230+ respondents).”
“Virtual and physical appliances are just a bundled operating system and application, but without any of the normal controls over system selection, build, hardening or management that is standard in a normal operating environment,” Mann explains. “Very few, if any, enterprises will allow any regular system -- even a radically slimmed-down system -- into their production environment without management capabilities like performance monitoring, patch distribution, firewall, malware detection, security administration, etc.” Until virtual system management appliances have these capabilities, he adds, “their uptake will be limited, especially in large enterprise environments with strict standards for new system deployments.”
Blomeier says he’d like to see VKernel add features like automatic baselining and templates for setting alerts, but that for his purposes -- continuously monitoring shared resources, receiving “proactive alerts” and being able to adjust capacity before trouble occurs -- VKernel’s appliance “is what we were looking for.”
Blomeier isn’t the only one looking for something more. In a survey conducted by Enterprise Strategy Group, only 24 percent of VMware users said they feel “very confident that the current tools for managing their VMware environment were sufficient to maintain existing IT service levels.”
Virtual appliances don’t cure everything, but that is not stopping their proliferation or significance. VMware offers a shopful of them, as does Sun, and some of the most prominent enablers of on-demand computing use appliances for one thing or another. 3tera, for example, runs applications as scalable appliances within its AppLogic grid operating platform. rPath offers a platform, rBuilder, for assembling virtual appliances for VMware and other environments.
May 16, 2013 |
When it comes to cloud, long distances mean unacceptably high latencies. Researchers from the University of Bonn in Germany examined those latency issues of doing CFD modeling in the cloud by utilizing a common CFD and its utilization in HPC instance types including both CPU and GPU cores of Amazon EC2.
May 10, 2013 |
Australian visual effects company, Animal Logic, is considering a move to the public cloud.
May 10, 2013 |
Program provides cash awards up to $10,000 for the best open-source end-user applications deployed on 100G network.
May 08, 2013 |
For engineers looking to leverage high-performance computing, the accessibility of a cloud-based approach is a powerful draw, but there are costs that may not be readily apparent.
05/10/2013 | Cleversafe, Cray, DDN, NetApp, & Panasas | From Wall Street to Hollywood, drug discovery to homeland security, companies and organizations of all sizes and stripes are coming face to face with the challenges – and opportunities – afforded by Big Data. Before anyone can utilize these extraordinary data repositories, however, they must first harness and manage their data stores, and do so utilizing technologies that underscore affordability, security, and scalability.
04/02/2012 | AMD | Developers today are just beginning to explore the potential of heterogeneous computing, but the potential for this new paradigm is huge. This brief article reviews how the technology might impact a range of application development areas, including client experiences and cloud-based data management. As platforms like OpenCL continue to evolve, the benefits of heterogeneous computing will become even more accessible. Use this quick article to jump-start your own thinking on heterogeneous computing.