As federal technology executives gain experience using virtualization technology to reduce the number of physical servers eating up space and power in their datacenters, many are starting to discover that virtualization can also offer similar efficiency and cost-cutting benefits for their business continuity capabilities.
Mike Rosier, senior systems administrator at Fermi National Accelerator Laboratory, explains how he and his colleagues are using virtualization to create a more resilient IT infrastructure for the lab for a fraction of the cost of traditional business continuity options.
Federal Computer Week: Can you give provide an overview of the general use of server virtualization at your organization?
Mike Rosier: At Fermilab, we've been using modern server virtualization technologies for over 5 years. In fact, I'm sure we were utilizing earlier implementations back in our mainframe days.
Some of the early reasons we decided to invest in virtualization were to address power and cooling issues in our computer rooms. This was at a time we were trying to keep up with the growing demands for development and test systems. The procurement costs for dedicated physical servers were also eating into our yearly budgets.
At this point in time, we've migrated between 60 and 70 percent of the physical systems that we originally identified as being good candidates for virtualization. As virtualization technologies continue to improve, we'll look to identify even more systems that may have originally been excluded from our list.
We’re supporting a wide variety of systems as virtual machines, including those used for test, development, integration, and production environments. Some of these include web servers, file servers, custom application servers, data acquisition systems, email servers, monitoring systems, authentication systems, terminal servers, and print servers.
In recent years, we’ve significantly reduced the number of new physical system purchases and now consider virtualization as a first option. Although it might not be a perfect match every time, we’re seeing fewer systems that do not make sense to setup as virtual machines.
FCW: Was using virtualization to support business continuity objectives part of the original impetus and business case for server virtualization, or was it a second stage objective?
Rosier: For the most part, using virtualization to support business continuity objectives has been a second stage objective until recently. While our virtual infrastructure continues to grow and mature, we see more and more customers looking to virtualization as a way to avoid costly clustering solutions, which can provide quick restoration of service after hardware failure or data loss. As service providers have become more and more aware of the capabilities virtualization can provide, we’ve spent just as much time discussing backup/replication options and failover strategies as we spend discussing virtual machine sizing and application specific requirements.
FCW: Can you describe with a little more technical detail how virtualization supports business continuity objectives?
Rosier: In order to describe how virtualization supports business continuity objectives, it helps to understand just what those objectives are in your environment. Some customers require their systems to be available 24x7, while others are satisfied with 8x5. Not only do systems need to be available, but they also need to perform adequately.
Virtualization can allow you to meet your objectives by allowing you to focus more of your efforts on configuring a relatively low number of redundant systems capable of providing enough failover capacity to weather varying types of outages. In our environment, we’re using technologies such as [network interface controller] teaming, redundant storage adapters and paths, live virtual machine migration, full virtual machine image and file-level backups, and cloning/replication. We’re also utilizing multiple data centers using separate power/cooling feeds to meet our business continuity objectives.
Virtualization has given us the ability migrate workloads from one building to another without impacting production operations.
FCW: How does using virtualization for business continuity compare technically and cost-wise to prior approaches for achieving availability objectives?
Rosier: When you compare the use of virtualization technologies to prior approaches for achieving availability objectives, you'll quickly notice how simple it can be to achieve server, storage and network redundancy. With today's technology, you can also easily achieve data center redundancy using fairly low cost solutions compared to what was available in the past.
Some of the earlier solutions providing business continuity required costly clustering software or hardware, and specific knowledge of how each of those solutions functioned in order to quickly recover a workload onto a different system. With the advancement of virtualization technologies, it becomes easier to provide the ability to recover from a hardware failure or to separate certain virtual machines from each other onto different physical servers. Less complexity generally translates into greater reliability.
Virtualization allows you to achieve high availability for a greater number of systems for a fraction of the cost if you consider what it might take to provide “like” hardware for each of your systems. Since the physical server hardware is often abstracted from the guest operating systems, most virtualization platforms make it easy to automatically restart or keep a guest running after a full system failure.
In some cases, we've been able to take a bit of a hybrid approach to providing business continuity for key applications. Mixing virtual machines and dedicated physical servers into a single cluster can be an option that not only saves on hardware costs, but also gives you a foot in each door if vendor certification is a concern. In many cases, devices such as load balancers can certainly function for both virtual and physical systems belonging to the same cluster.
FCW: What pitfalls would you warn others to avoid when using virtualization to support continuity objectives?
Rosier: You should always be prepared to discuss your strengths and weaknesses. If you find shortcomings in your environment, be sure to investigate, identify, and prioritize cost-effective solutions that can be easily integrated into your virtual infrastructure. Virtualization technologies continue to evolve rapidly, so make sure you keep up with industry trends before making any large, strategic purchases.
Make sure your customers and management chain has a clear understanding of your capabilities are and what you're protected against when failures occur. For example, high availability means different things to different people. Sometimes it means no downtime and sometimes it means minimal downtime. Make sure this is clear up front. You should simulate failures and test your resiliency from time to time. You certainly don't want to find out that you're not protected against something you invested in heavily to avoid. It could reflect poorly on you and by extension, your organization.
Depending on the size and structure of your organization, you may need to engage members of other groups to help you meet your business continuity objectives. Don't miss this step! Unless you have a clear understanding of all the resources your virtual infrastructure is dependent on or interacts with, there's a good chance you'll make assumptions that can cost you time and money in the future.
For example, if you purchased iSCSI storage arrays as a cost-effective way to provide storage to a new data center, you might soon learn that there are network switches in your path that do not support or are not configured to support jumbo frames, which can be a requirement for certain workloads. Or, maybe you discover that your fibre channel switches and servers might support 8Gb connections, but your fiber [cables] support less than half of that in a reliable manner.
As they say, “The devil is in the details,” so try to learn those details by communicating with the experts in each area before you purchase and deploy a technology intended to address business continuity objectives.
Posted on Jan 20, 2012 at 12:28 PM1 comments
There’s a growing wind in the sails of microservers, a new type of datacenter computer that is extremely energy efficient and tailor made for the cloud- and Internet-style workloads that are increasingly common at agencies and other big enterprises.
Dell joins some smaller players and recently introduced its initial line of microservers, Intel has begun shipping the first of several processors designed specifically for use in microservers, and Facebook officials say they have big plans for the diminutive computers.
In a recent story we covered the reasons why microservers are expected to make a huge splash even though they buck one of the hottest trends in enterprise computing, the use of virtualization software to consolidate data processing chores on to fewer, more powerful servers. If you are interested in learning more about microservers, read on for some links to key news items, case studies, analysis and technical discussions.
Microservers put a low-power microprocessor, like those designed for smartphones and tablets, on a single circuit board, then pack dozens or hundreds of those cards into one server cabinet that provides centralized power supply, cooling fans and network connections.
There are many opportunities for using microservers to drastically reduce datacenter operating costs, including:
* Web applications with high volumes of small discrete transactions, like user logins, searches, checking e-mail and simple Web page views.
* Running hosting services that provide a dedicated physical machine for each application or user.
* Creating grids or clusters in which multiple server nodes work in parallel on a specific task.
* Environments that need to reduce the energy consumption and physical footprint of their data center servers.
Some folks think microservers will eventually dominate most cloud datacenters. John Treadway, global director, cloud computing solutions at Unisys, makes a persuasive case for microservers on his personal blog CloudBzz. He predicts that microservers will replace bigger servers running virtualization in most commercial cloud datacenters by 2018, with internal enterprise datacenters on the same path though a few years later.
To see what a large-scale cloud datacenter packed to the gills with microservers looks like, click on the YouTube video available on this webpage from Data Center Knowledge. It’s taken inside French hosting company Online.net’s facility and features early microservers built by Dell. The section showing the microservers begins about 1:20 into the video.
Facebook is one of the highest profile players in the U.S. to endorse the microserver approach for large scale data centers. In this article from PCWorld, Gio Coglitore, director of Facebook labs, lays out the rationale for the social networking giant’s plans to move to microservers, with reasons that include energy efficiency, avoiding virtualization vendor lock in and increasing system resiliency.
One of the better write ups that clearly explains the value proposition for microservers is available from one of the industry’s earliest players, SeaMicro. It’s a case study about Mozilla, the group that organizes the development of the Firefox Web browser, and their use of microservers. One of the most interesting parts of the article describes how Mozilla officials calculated, among other cost and efficiency metrics, the energy required to perform a certain processing task, in this case an Internet HTTP request. They concluded that microservers used one-fifth the power per HTTP request than a traditional server.
Correlating power consumption with the work output of an IT system is a more advanced and meaningful way to calculate datacenter energy efficiency than the metrics CIOs now use most often. Last year I wrote a story about efforts to increase the use of these more sophisticated metrics. [
If you really want to get into the weeds about the relative performance of different processor approaches and their suitability for varying types of workloads, there are a couple of good papers that size up those debates.
One paper from a group of researchers out of Carnegie Mellon University evaluates the use of clusters of low power chips like in microservers deployed in what is called a Fast Array of Wimpy Nodes. The FAWN approach can be a much more energy efficient option for many types of workloads, but not all. The researchers note that the power costs of large datacenters accounts for up to half of the three-year total cost of owning a computer.
On the other hand, Google released a paper from one of its researchers that details the drawbacks in certain situations of arrays of wimpy chips. What happens is that wimpy-core systems can require software applications to be specially written to run on them, resulting in extra development costs that can take a big bite out of the energy savings.
Posted on May 27, 2011 at 12:07 PM0 comments
We’re working on a story for FCW about the Department of Veterans Affairs “Blue Button” Web application that allows veterans to download their personal health information from the department’s MyHealtheVet site.
We’re looking for veterans who have used the VA’s Blue Button to share their opinions about the application’s usefulness with the FCW community. If you’ve given this application a try, you can use the Comment button below to tell us about your experience.
One veteran who checked out the application and wrote about the experience on a blog last fall reported being distinctly underwhelmed by the experience. “Here’s the cream that floats to the top, the icing on this cake, the best of the best; If you download and install the Blue Button to your personal computer, you will be able to securely access and download and print and share all the data that you yourself put in to the system,” wrote Jim Strickland, a veterans’ advocate.
Since then, the Centers for Medicare and Medicaid Services launched its own Blue Button application on its MyMedicare.gov website. That feature lets 47 million Medicare beneficiaries view, download and print their medical records.
Posted on Apr 08, 2011 at 1:25 PM0 comments
Finally, thanks in no small part to the recent Telework Enhancement Act, it looks like a lot more government offices will be giving telework a try. Previously resistant managers are coming on board (for the moment, anyway), identifying positions for telework eligibility, dealing with equipment needs, and developing agreements about employee performance and expectations.
Of course, most telework programs start with the best of intentions, but not all march on to meet great success. Employees who abuse the telework privilege with lackluster performance hurt productivity and can poison office morale. They can also jeopardize management support for the telework program. Sometimes managers have to work with poorly conceived telework policies, so they lack tools that could help them bring wayward employees in line.
So, tell us, what are some of the mistakes that employees or their managers can make with telework? And what can they do to avoid falling into the same traps?
Readers who have commented on past FCW stories about telework have mentioned some of the problems that can arise.
One mistake mentioned is the teleworker who isn’t responsive to communications from managers and co-workers. “I have been constantly frustrated and so have others in my division when you try to contact (phone/email) a teleworker for immediate answers/assistance and they do not respond quickly,” wrote one reader.
And what can managers do for their part to make telework a success? One reader said managers should give telework a chance but should also be ready to rein it in if it’s not working. “A manager must immediately send a clear signal and not give so much as 1/64th of an inch but yank the privilege upon the slightest infraction. When employees know that they will not be allowed to take advantage of a given situation, they quickly fall in line.”
But some managers don’t have this kind of power. They talk about being hamstrung by policies that don’t allow them to revoke telework privileges, even when some employees are clearly abusing them. “As a manager I should have the authority to approve or deny telework should there be an upcoming holiday,” writes one reader. “Employees go through the calendar and always telework prior to and following a holiday — [that] should not be allowed.”
So what do you think? What are the problems that can undermine telework programs, and how can they be avoided? What kinds of policies do managers need to make telework successful? Please share your comments below.
Posted on Mar 11, 2011 at 11:16 AM2 comments
E-mail is the hands-down winner for the most popular high profile cloud application that government agencies are moving to first. The General Services Administration, Agriculture Department, Interior Department, Lawrence Berkeley National Laboratory, California and Wyoming, and New York City and Los Angeles are among the government entities and jurisdictions that have already moved or will move their internally hosted electronic messaging systems to a commercial cloud e-mail service.
Experts whom I talked to for a story on cloud adoption trends expect this pattern to continue, as e-mail represents a prime opportunity for tapping the benefits of a commercial cloud service.
Most agencies wrestle with managing multiple legacy e-mail systems that are usually expensive to maintain and incompatible with one another. That situation affects the productivity of the IT department and agency end users. Moving to a single enterprise-wide e-mail system makes a lot of sense. And the commercial cloud providers are adequately dealing with security and reliability concerns to enable government to make the move.
So what do you think the next hot cloud application will be? What cloud services is your agency most interested in adopting first and why? You can post your comments below.
Also, which types of cloud services do you think should be the easiest to adopt, relatively speaking? Alternatively, which cloud services will provide the most benefits or biggest payback to your agency, even though they may not be the easiest or quickest to adopt?
Shawn McCarthy, research director at IDC Government Insights, expects some other software-as-a-service (SaaS) applications to follow e-mail as lead targets for early cloud adoption. The universality of certain applications across government creates an opportunity for cloud vendors to create standard services that can meet most agencies’ needs, he said.
Like e-mail, human resources systems, document management and certain types of financial transactions all tend to be common functions that can be standardized and potentially moved to a more cost-effective cloud model, McCarthy said.
In the category of infrastructure-as-a-service (IaaS), he said cloud storage services show strong early promise.
But even these applications are not slam dunks for the cloud at every agency. For example, existing in-house e-mail systems are sometimes integrated with other applications such as case management, contact management or systems that track legal document delivery. That kind of integration can make a move to the cloud more complicated, McCarthy said.
Other factors related to the different characteristics of government agency’s also influence whether they use the cloud and which style of cloud applications they choose, said Greg Potter, a research analyst at In-Stat and author of a recent report about the commercial or public cloud market.
According to his research, small cities and counties, just like small businesses, lead in the adoption of public cloud services, because these organizations traditionally outsource their IT needs more often than larger government enterprises.
Also, SaaS accounts for more than half of the $130 million government spent on public cloud services in 2010, Potter found. Government spending on IaaS in 2010 was roughly $40 million and only $18 million for platform-as-a-service. He estimates the government spending on public cloud services will increase to $275 million by 2014.
Potter is starting some research on private clouds, in which organizations build their own cloud-based services for internal use. This approach eliminates the data security and control concerns many organizations have about using commercial public cloud services.
Potter expects to find that government spending on private clouds will be significantly higher than that on public cloud services. Moreover, he expects the profile of who’s spending most on private clouds to flip, whereby the majority of that spending will be by large- and medium-sized government entities, those who traditionally have had the resources to take care of their own IT needs.
Posted on Feb 09, 2011 at 9:32 AM0 comments
Mobile devices like smartphones and tablet computers are the hottest of the hot gadgets these days. In the case of smart phones, at least, many agencies dole them out to some employees as standard-issue equipment for voice and mobile messaging needs. But in other cases, government employees are using their own personal devices, from Android smartphones to Apple iPads, to check in and stay on top of their office chores.
However it is that these devices get into workers’ hands, this “consumerization of IT,” as it’s called, has many agency chief information on notice. They understand that the devices are going to reshape how government work gets done, but they also know that the gadgets pose new management and security challenges that have to be addressed.
What are your agency’s policies about smartphones and handheld computing devices? Are there certain devices that your agency issues and supports? Are there others that they don’t but you wished they did?
Also, what kind of workplace role do you see these devices playing? And what will be the biggest stumbling blocks in the way of this happening?
Please share your comments below.
Posted on Nov 09, 2010 at 11:23 AM3 comments