Datacenter consolidation refers to the reduction of disparate locations from which any organization hosts key applications and services. A well-executed datacenter consolidation project can deliver many significant benefits. Fewer locations set the stage for:
- A reduction in network equipment and the number of servers required, which in turn paves the way for the introduction of application virtualization; (In fact, a 2008 EMA survey estimates nearly three-quarters of all enterprises employ virtualization for at least some of their production applications.)
- A streamlined transport infrastructure amongst remaining sites
- A more secure, compliance-driven architecture, with fewer physical facilities to oversee and fewer conduits over which business-critical information must flow
As a result, datacenter consolidation offers the potential to slash capital and operating expenditures, improve productivity and service levels, and mitigate and minimize risk.
Key Issues
While the advantages of datacenter consolidation are clear and compelling, what’s also evident is that the performance of every single resource is absolutely critical to success. Network equipment and servers must be available at all times. Applications must not only be available, but must provide lightning-fast response times to their users. And the proliferation of technologies like multi-protocol label switching, class of service, and quality of service, coupled with the convergence of voice, data, and video, means the transport is no longer just a connection between two points. In a consolidated world, it must be an integrated part of an overall application aware infrastructure that will be relied upon to fulfill the promise of improving service levels without sacrificing security.
If business units are to relinquish control of these assets as part of the datacenter consolidation effort, they will expect something in return: improved performance to assist them in the execution of their responsibilities, and accountability to ensure they are charged only for the resources they use and to demonstrate the achievement of service level metrics. To meet these objectives, deployment of performance management solutions is essential.
Data Center Consolidation can help to maximize assets, but a holistic approach to taking on consolidation is necessary to minimize impact on the business during the transition.
Approaches
Most companies possess a variety of performance management solutions that support their distributed architecture, and these capabilities typically are collected and redeployed together as part of the datacenter consolidation effort. These disparate tools tend to operate in silos and have unique user interfaces and data models, which makes them nearly impossible to integrate cohesively. As a consequence, operations staff must conduct a “swivel chair” style of management that hinders performance monitoring and extends the problem identification/isolation/resolution process.
In addition, legacy solutions simply aren’t comprehensive in the depth and breadth of information they collect. In terms of depth, some tools focus only on high-level performance data, which is presented via dashboards for a quick overview. Other tools drill down to the level of packet capture to provide detailed analytics. Unfortunately, there is no way to marry these two categories of information, which is imperative to root cause analysis. With regard to breadth, most existing solutions either don’t collect all available information (limiting themselves to only the “Top N” items) or don’t store the data for a long enough duration. Both are critical to the identification/isolation/resolution of more subtle, complex, intermittent anomalies that often plague performance.
Best Practices
The most successful datacenter consolidation projects do more than merely bring network equipment, servers, and applications together under one roof. They leverage the opportunity to truly redefine the enterprise architecture. A similar sentiment and mandate is necessary with respect to performance management. What’s required is a next generation solution that recasts performance management to best meet the needs of the consolidated datacenter environment by:
- Collecting information from all possible sources, including appliances, dedicated probes, client and site agents, polling, and infrastructure data;
- Converting all information gathered into a common data model that tightly binds the information together, regardless of source;
- Archiving and maintaining all information over extended time periods to support the baselining of nominal performance over the short-, medium-, and long-term;
- Presenting real-time and historical information with hierarchical displays and reports that enable “sweeping and swooping” from summary dashboards all the way down to individual transactions; and
- Providing metrics and views that depict quality of experience from the end user’s perspective.
Fluke Networks’ Visual Performance Manager is the most comprehensive next generation performance management solution available today. Its network performance appliances provide network performance and usage views by interacting with routers, switches, and other network equipment to obtain the detailed netflow information generated by these devices, regardless of format. Its application performance appliances rely on patented, proprietary technology to capture, filter, and save inline or spanned data associated with every application-based transaction – even for those applications that have been virtualized. And its analysis service elements deliver the Layer 1-7 visibility into the wide area network crucial when more stringent performance criteria, such as service level agreements and network, service, and application availability, are in play.
With Visual Performance Manager, performance problems can be identified, isolated, and resolved with unprecedented speed and accuracy – eliminating a significant obstacle to datacenter consolidation by mitigating the risks from planned changes and unexpected events. The solution’s open architecture means its components, displays, and reports can be integrated and configured to meet an organization’s precise requirements. Its Web-based, fully URL controllable design gives companies the option to integrate some or all of their legacy tools into the platform to leverage their prior investments – helping the datacenter consolidation project meet or exceed all of its cost and performance goals.