This vendor-written tech primer has been edited by Network World to eliminate product promotion, but readers should note it will likely favor the submitter’s approach.
Many organizations have been consolidating data center resources to drive efficiency and save money, and some are now poised to continue that push in branch locations using converged infrastructure to combine compute resources, network control, and storage delivery in a single scale-out unit that reduces hardware and eases management and maintenance.
Converged infrastructure is designed with virtualization in mind and optimized for quickly provisioning infrastructure capabilities to support apps, developers or other IT teams. One bank was able to reduce the time it takes to provision a virtual machine for a developer from six weeks to one hour. They will soon have it down to six minutes. It’s important to create an application-tuned infrastructure. The other benefit is how you can configure the converged stack to optimally serve up various applications.
Converged branch office infrastructure also reduces costs because you don’t do the research and development. Instead, the integration is done by the component vendors, allowing data center teams to deploy the technology significantly faster. Better yet, you receive a single support number to troubleshoot any part of the converged solution.
However, converged branch office infrastructure won’t be completely the same as converged infrastructure in your data center. Why would it? You have a unique set of needs in branch offices, which by definition are further from centralized IT, but closer to end-users.
You can’t modernize your branch office without taking user performance experience into account. If you’re not careful about how you consolidate your branch infrastructure, performance will suffer. And that’s likely to hamper productivity and frustrate branch office workers. A major benefit to branch office infrastructure convergence is to make applications easier to manage and more reliable for end users.
Recognizing that the branch office is the next frontier of data center transformation, but understanding the importance of performance, we recommend the following capabilities for your branch office converged infrastructure.
* Virtualization with central management: Virtualizing servers is a given in converged infrastructure, but automation is critical, especially if your company has hundreds of virtual machines to track. Using tools like VMware’s vCenter to help manage and optimize virtualized instances is a key ingredient when introducing virtualization to the branch office.
* WAN acceleration: While virtualization in the branch is nice for local services (like print, DNS, etc.), the reality is most applications will be most efficient back in the data center. Unfortunately that means branch workers are far away from many of the apps and data they need to use. Yet they need rapid and reliable access to information to maintain productivity. WAN optimization can make this happen by reducing the chattiness and inefficiencies of application protocols and TCP. This enables more consistent, faster experiences for users no matter where they are working.
* Troubleshooting support: When issues arise and there’s no local IT staff, that’s a problem. You need a way to capture and analyze packets remotely through the converged infrastructure. With monitoring support built into converged branch infrastructure, IT can see right away when performance at the edge is suffering.
* Storage delivery: The big gotcha with centralized storage is that user experience at remote offices often suffers. Storage delivery technology addresses this problem by enabling fast access to centralized data and eliminating the need for branch backups. Storage delivery acts like a vacuum cleaner on a company’s storage units in the data center, pulling data out to the branch quickly and pushing local updates back to the SAN. In the event of network outages, storage delivery units maintain some data locally. Once the network is back online, a quick reboot of the branch server restores connections to the centralized data.
* Stream splitting: If your company is a frequent user of video streaming, for company-wide meetings or mandated training for instance, look into stream splitting to lighten the load on the network. Research shows that business video traffic may double in the next few years. Don’t leave your branch locations out of the loop because your network can’t handle video streaming over long distances.
* Quality of Service (QoS): VoIP, Web apps, video streaming, and enterprise database applications all have different requirements and place unique demands on the network. To ensure a predictable user experience and prioritize infrastructure resources according an application’s importance, you’ll need the ability to classify applications and allocate bandwidth automatically with predefined policies. QoS helps deliver consistent performance to branch users while ensuring high availability, which is particularly important for business-critical and latency-sensitive applications, like ERP systems, voice, and VDI.
* Network Path selection: With the accelerated adoption of hybrid networks globally, IT must be able to effectively route traffic across different public and private, MPLS networks. Network Path selection technology enables IT administrators to cost effectively and reliably manage application performance across hybrid networks. This requires classifying applications according to business priority and resource consumption, such as bandwidth use and defining the optimal network path for applications including backup paths for high availability. Path selection technology also monitors the network connections for performance and provides failover mechanisms when network links are down.
That’s a long list of things to look for in branch converged infrastructure, but the ever-increasing diversity of applications and networks requires more sophisticated capabilities at the edge of networks.
The common theme here is striking a balance between making it really simple and cost-effective to operate IT in remote locations, while still providing users with the performance that they need.
Raj Mallempati is VP at Riverbed