 |
|||||||||
|
|||||||||
|
|||||||||
According to a new report by McKinsey & Company and the Uptime Institute, there's too little demand and capacity planning within IT management. That translates to 6% average server utilization and 56% facility utilization.
Data center growth is beginning to cause significant strain on IT budgets and will continue to do so as companies add more servers. For every $1 spent on hardware, $8 are spent on maintenance and $.50 is spent on heating and cooling. Server over-provisioning is rampant, and companies are looking for safe ways to consolidate.
Many companies are turning to virtualization as they look for ways to consolidate servers and reclaim precious data center space. By consolidating a data center's physical servers into virtualized servers, the enterprise can realize higher utilization rates and increased operational efficiency. The result is reduced costs associated with software, maintenance, power, heat and air, ultimately lowering TCO.
Virtualization Transformation
With the increase in economic and environmental pressures, the use of virtualization is on the rise. According to IDC, 22% of servers are virtualized today and that number is expected to reach 45% in the next 12 months. Using virtualization to consolidate servers requires careful planning, sizing, and configuration. Improper sizing or faulty configuration of a virtualization platform can negatively affect performance or uptime of numerous virtual machines (VM's) rather than just a single instance.
In terms of application performance in virtual environments, you should consider all aspects of the data center infrastructure; networks, servers and storage all affect performance. One of the biggest challenges you will face is related to VM configuration on your hardware. The question becomes, what mix of applications can safely be consolidated onto which virtual machines without breaching Service Level Agreements (SLA's) for performance and availability?
Decision Support for Physical to Virtual Migrations & Consolidations
Application consolidation may be a relatively simple task when there are only 10 servers in the landscape. But when hundreds of servers need to be consolidated, using a "Best Guess" approach is not good enough. Combining applications that have similar peak transaction times, for instance, could have serious consequences, resulting in unnecessary downtime, missed SLA's and consequent funding challenges. The predictive modeling capabilities of HyPerformix allow you to compare unlimited consolidation scenarios to determine the optimal placement of individual workloads in your virtualized environment. You can evaluate multi-tier applications crossing heterogeneous physical and virtual environments. All this is done prior to actually making any changes, so you can be sure you're making informed decisions.
Virtual to Virtual Capacity Analysis & Planning
You will face even more challenges once you have deployed virtualization. Traditional monitoring agents look at an entire server, not the individual VM's, and thus, the metrics reported from the agent can end up being largely useless. Using HyPerformix, companies running SAP on IBM AIX LPAR's are able to combine both SAP transactional information and AIX virtual metrics to do predictive capacity planning for physical and virtual environments.
The resulting model can then be used to:
• Determine host and LPAR utilizations
• Identify LPAR's that exceed entitlement
• Evaluate scalability and workload growth scenarios
• Model virtual configuration changes to ensure performance and response time SLA's
• Simulate instance consolidations
• Model NetWeaver Upgrades including hardware refreshes to, and between, virtual platforms or configurations
• Generate specific recommendations for IT investment, deployment and configuration
Accurate reporting of virtualized metrics correlated with your business applications can be a difficult task. Things get even more complex when you consider that each virtualization vendor's hypervisor (VMware, AIX, HP-VM, SUN, etc.) have different performance capabilities and overhead. For companies running SAP, the native SAP metrics only provide data on the SAP application itself, not the overhead consumed on the server by the hypervisor. To get the entire picture you must be able to see the total resources of the server in addition to the VM's that are running on it. HyPerformix leverages detailed server and virtual machine utilization metrics, including such factors as virtual CPU usage and entitlements, for accurate analysis.
Data Sheets
White Papers
10 Questions about Virtualization
Encouraging a Simpler Path to Green IT
Taking the Risk out of Server Consolidation & Virtualization
Successful Virtualization in the SAP Landscape
Solution Briefs
Webcasts
Partnering with IT to Manage Performance with Forrester Research
Product Reviews
Presentations
Transformation to Virtualization
Using Effective Performance Management to Improve Service and Reduce Cost
"Consolidation is a transition from an unpredictable, inefficient state to one that is sustainable and efficient. On the front end, economic analyses like reduction in operational costs and percentage of IT budget saved can be very effective in making the case for investment to C-level execs... economic metrics will always be critical to securing future support and funding from the executive team."
James Staten, Forrester
