1. VIRTUALIZATION—SERVER CONSOLIDATION

According to VMware, only about 15 percent of servers have been virtualized worldwide. So that leaves 85 percent of servers left out there. How can we fast track virtualization in the data center? It’s not that companies haven’t bought or made investments, it’s that customers either adopt the technology too little, too late—or too much, too soon.

Some companies have invested in the virtualization process but their ability to leverage virtualization is limited by an older physical environment. Others may have the infrastructure in place but aren’t taking advantage of all the features native to the technology. Some examples of this include customers that just used virtualization to contain their new rollouts, but stalled when trying to consolidate existing servers.

Some Reasons to Fast Track Virtualization

• Free P2V tools are readily available. For example, a customer may have VMware in place but isn’t necessarily taking advantage of Virtual Center and Converter to completely turn their physical environment into a virtual one.


• Average savings per 200 servers virtualized is $200k in power and cooling costs alone each year! Add in the average cost of more than $7000 per server in just hardware, and customers can remove approximately $1M of hardware from their data centers!


• Today, 95 percent of applications are fully supported by the vendors. Three years ago, this wasn’t the case. VMware is driving the industry to support 100% of the core business applications and this is becoming reality. This means that more customers can now take business critical applications and virtualize them, all while improving DR, resource management, flexibility and agility in their infrastructure.

2. RESOURCE POOLING

So how can you accelerate your virtualization plan? We don’t see a lot of people taking advantage of the fact that you can take your whole infrastructure, cluster it into one cohesive unit, provision off, and manage it as one big object. In fact more than 45 percent of customers haven’t used resource pools or clusters or VMotion which are some key features you can leverage once you get past the point of server consolidation.

Another consideration is people looking at their environment in terms of VMs rather than applications. The ability to think in terms of applications is a major mind shift. By being able to consolidate more applications per physical server, cost per application is drastically reduced.

3. STORAGE TIERING

The great thing about storage tiering (especially with virtualization) is that it allows you to create levels of storage based on types of probabilities and cost considerations in relation to running your business. For example, you may have a critical app that requires a high level of storage performance, and therefore cost isn’t a huge consideration since it’s critical to your business. Or you may have a run of the mill portal that may not need to be at a high level of performance. Using something such as Storage VMotion, you can move a vm between these different priority levels as your needs change in regard to the application. Storage can then be purchased based on availability and performance needs to the ever-changing business priorities and re-prioritized on demand.

4. THE CASE FOR 10GB

70 percent of all IT costs go into the maintenance of applications and the environment, which means there’s not a whole lot left over for buying new technology. So if you simplify, it can make a huge difference. For example, the ability to collapse your infrastructure and provision heavyweight storage protocols (FCoE, iSCSI, NFS) over one big backbone is a tremendous advantage. Cisco and VMware for example are working to create virtualization in the data center in a platform that covers storage, networking, virtualization, and the management to handle it all.

Which bring us to the fact that 10 Gigabit Ethernet delivers 10x the bandwidth, and some switches and NICs implement additional features that support consolidation of multiple physical links onto one, including the capability to create lossless priority classes for iSCSI, NFS, and FCoE traffic. Plus, FCoE is a straightforward encapsulation of Fibre Channel into Ethernet that preserves the lossless character and management models of native Fibre Channel networks.

FCoE will be supported on most 10 Gigabit NICs as well as by a new class of converged network adapters that appear to the OS as separate Fibre Channel HBAs and NICs, making FCoE’s existence completely transparent to the operating system.

5. PATCH MANAGEMENT

This is always a sticky topic. Microsoft puts out 79 patches a year and with an average of 12 minutes per machine to do each patch (according to Gartner) once you add it up, it becomes 175k a year per 100 machines for automating patch management. So the question sometimes becomes: to patch or not to patch. The trend change around virtualization is that rather than patching each and every machine, we’re patching at a level below the hardware level. This allows us to push out a patch once and patch everything in the infrastructure—and the OS doesn’t know that it was patched.

VMware provides a great tool called VMware Update Manager that leverages Shavlik technology to patch online and offline virtual machines. The advantage here is Shavlik also has a strong solution for patching and maintaining physical environments as well so IT organizations can manage patching from one pane of glass.

6. ARCHIVING LEGACY DATA

One of the things to remember with archiving is that these technologies require companies to make some critical business decisions around how to prioritize data retention and define policies. With all of the advances in archiving and backup, the largest pain point to success remains defining the business requirements, not the functionality of the technology. We see a lot of companies backing up data that will never be looked at again. In fact, more than 85 percent of backup data is never viewed again. The key here is to prioritize because not all data is created equal. What do I need to keep and how long?

Get down and ask the business units these questions and most organizations will be surprised with how many cost saving opportunities are uncovered. Maybe you find a tape archiving unit you don’t need, or better yet, maybe you don’t need tape as a medium at all. Perhaps MS Exchange or SQL databases in your environment don’t need every last piece of data backed up on a nightly basis since it’s already stored on near-line disk. GreenPages Professional Services specializes in undertakings like this—it’s more than just implementing technology; it’s digging down deeper and prioritizing what really matters to the business and creating a cost effective plan.

7. APPLICATION VIRTUALIZATION

Let’s face it, building, rolling out, managing, and patching applications is an IT operations nightmare. Some customers have 10—100 different images rolled out across their infrastructure leading to configuration drift. How can an organization begin to be standardized with this many moving, changing variables?

By packaging applications into a virtualized application, you separate the app from the OS which means NO MORE registry edits, shared DLLs, or admin rights on end user systems! By virtualizing the application, you can build the permissions inside the application, and when you roll it out you don’t have to give full admin access to the machine; this is a huge security benefit. Virtualizing apps with VMware ThinApp also allows you to more easily move to new desktop platforms; apps are no longer tied to specific versions of Windows and upgrades become much easier. Imagine being able to create 1-2 desktop images and only provisioning applications to users that need them based on groups!

8. LEVERAGING HIGH AVAILABILITY

Many near-business critical apps do not meet the business’ availability requirements or don’t require expensive failover solutions. The advantage of virtualization is that it not only allows you to consolidate servers, it also allows organizations to revisit apps that didn’t have a robust DR plan based on cost considerations. For example, maybe an application wasn’t critical to the overall business but it was critical to a business unit of your environment—the cost of providing necessary DR wasn’t justified by the smaller impact.

Less than 50 percent of VMware customers are using the HA (High Availability) feature that they already own. The great part is that as you go down the virtualization road, it’s easy to click the button and turn it on. One key thing to remember however is that apps are prioritized based on criticality. By understanding this key step, organizations can build resource pools around the HA cluster based on business units or criticality of function. GreenPages Professional Services is specialized to assess, plan, and design an effective, highly available infrastructure to meet the needs of the business.

9. REDUCING UNNEEDED BANDWIDTH

A lot of companies have applications that once did fine on their network, but over time with upgrades to newer versions, additional features such as IM, multimedia, and realtime streaming, networks are now no longer keeping up. Also organizations change—maybe a company now has 10 remote offices with users that need the same level of support as corporate with data backup, a reliable connection to apps, patching, etc. Supporting these remote offices uses valuable resources.

In addition to deploying virtual desktops which reduce overall bandwidth required at locations, WAN acceleration and caching (such as those features found in the Cisco WAAS products) can help by speeding up the connection by not resending the same data. Another name for this is network data de-duplication. For example, if I make a change to an Excel spreadsheet, rather than copying down all the bits of the file, it only copies the changes down. This can save you a ton of reduced bandwidth and increase performance as well.

10. POWER AND COOLING RIGHTSIZING

The average power and cooling costs for one 2-way server are approaching $1500 a year. Many CIOs have been sold on virtualization on the sheer merits of the ROI. But the ROI was based on the assumption that the unnecessary physical equipment would be removed from the data center. What we’re seeing is that in many cases it hasn’t been removed—which completely negates the ROI of power and cooling costs. Opportunity lost.

VMware has a Dynamic Power Management feature that allows you to smartly power down parts of your virtual environment as demand and resource needs go down. Picture a data center with 200 servers that reduces to 100 powered on servers for 12 hours at night—That’s a 25% annual reduction on top of all the money already saved by consolidating!

And with power comes heat—by reducing or shifting server workloads throughout the data center, it is possible to more efficiently cool the environment, preventing hotspots and air conditioning overkill. As this technology matures, (APC is one of the leaders in this space), it will be possible to shut down every other server in a rack as an example—or a whole side of your data center—to reduce the thermal and power footprint. Now that’s leveraging a highly flexible and agile infrastructure!

Where to Get Started

GreenPages offers a comprehensive data center assessment that identifies key next steps based on your business priorities. We make sure your priorities match up with technologies that you may have implemented to make sure you are taking the right next steps. The solution may not mean buying product, it might be taking advantage of products and solutions you may already have. So on behalf of GreenPages, I challenge you to address your 2009 IT needs (and beyond) by isolating what you have in your environment and looking at your business priorities. Still need help? GreenPages provides leadership in IT to help you make the right decisions for your organization.

For more detailed information on ways to improve your IT performance and unlock the hidden ROI in your infrastructure, view GreenPages’ April Webinar (presented by Brian Gagnon) here or call your GreenPages account manager at 800-989-2989.

‹ ‹ ‹ Back to Newsletter