To save more money, do you have to expand your virtualization program to include performance-sensitive business critical applications? It was easy to virtualize those less critical internal IT, dev, test, and mailroom applications. Unfortunately it's quite a risk to tell critical business application owners not to worry -- "Your new virtual server will always perform well!" -- and then live up to that assurance.

Virtualization does not guarantee performance

Virtualization saves money in many ways, but it doesn't inherently guarantee application performance. In fact, using virtualization to leverage the sharing of powerful resources will likely improve average performance for all applications most of the time, but can easily degrade performance for key applications at peak times. The possibility of a negative impact to the most important users conducting their most productive business prevents many companies from virtualizing more than their "low-hanging fruit."

Deliberate performance management is still required

In the time before virtualization, it was considered poor practice to neglect capacity planning and rely on "firefighting" techniques for dealing with user complaints of poor performance. Although virtualization gives an IT administrator faster methods of reacting to performance bottlenecks by dynamically allocating and shifting resources, this is still at heart a reactive practice and one depending on as much of the administrator's art as it does science.

Despite virtualization vendor assurances that technologies will soon appear that will react automatically to fully address performance problems, reactive-mode system management is just not good enough for critical business applications. You may have found that seat-of-the-pants performance management in virtual environments isn't even keeping your secondary applications out of deep and recurring trouble. Since few new performance issues stem from previously experienced root causes in the exact same ways, and most components and inter-relationships within a virtualized system are dynamically changing, simply automating past performance mitigation with reactive policy engines can quickly make a cascading mess of out whole clusters of servers affecting hundreds of applications when merely trying to solve a minor problem.

-1-

Explore Our Topics

• BPM & Enterprise Architecture
• BPM Standards
• BPM Implementation
• BPM Best Practices
• Continuous Process Improvement
• BPM Governance
• Service Governance

• Business to Business (B2B)
• BPM for Supply Chain
• BPM for Customer Management
• BPM for Financial Services
• BPM for Healthcare & Insurance
• BPM for Government

• Business Activity Monitoring
• Event Processing
• Process Intelligence
• Business Rules Management
• Decision Management
• Operational Intelligence

• Workflow Automation
• BPM & Enterprise 2.0
• BPM & Web 2.0
• BPM & ECM
• Case Management
• Collaboration
• Enterprise Portals
• Social BPM

• BPM & Cloud Integration
• BPM Analysis & Optimization
• Business Process Simulation
• Process Visibility & Monitoring
• Process Orchestration
• Process Discovery
• Process Modeling

Home
Webinars
Blogs
ebizQ Forum
All Topics
Subscriptions
White Papers
Analyst Corner
Follow us on Twitter
Fan us on Facebook

Explore Our Topics

BPM Strategy
BPM Solutions
Event Processing Strategy
Content-Centric & Collaborative BPM
Process-Centric BPM

About Us
Editorial Submissions
Feedback
FAQ
Advertise With Us
Site Map
Privacy Policy
Unsubscribe