If there ever was an easy reason to adopt Service Virtualization techniques, it would be the performance lab. We are seeing customers reckoning with this problem over and over: highly skilled performance testers, with expensive lab infrastructure and software, WAITING around to access needed resources to do their job and ensure expected software performance levels.
In a couple weeks we are doing a big webinar on IT Briefcase with our
partners Accenture: "Addressing a Lack of Capacity: Virtualizing the
Performance Lab for Optimal Results."
The free session is June 24, 2010 1:00 - 2:00 pm, Central Time (2-3PM Eastern)
I am presenting with Accenture Partner Scott Gilliland, who runs the Global Performance Engineering practice for the firm, with experts who have already implemented these tactics to some incredible results at our joint customer sites - we are talking increasing the amount of performance testing done to go well beyond 6-Sigma reliability, at less 25% of the existing labor and infrastructure cost.
Today's applications are more distributed and complex than ever - and that makes for an entirely new economy around needing high performance that can't be ignored. Fail to perform, and you will certainly lose customers. Scott will expertly describe the financial reasons for eliminating constraints from performance testing.
What's interesting to note is that we usually find that performance labs are acutely aware they have idle time waiting for assets to run their load tests against, and wasted effort on coding simulations. But they don't know what these inefficiencies actually cost the business - until now.
Short of building a copy of the whole enterprise's software infrastructure for each test lab (which can cost millions for a typical F1000 enterprise), and manually coding all of the dependencies (such as mainframes owned by other divisions or partners, SaaS providers, Cloud-based services) needed to respond to the performance tests, there is another way to achieve a realistic environment that is truly elastic.
Service Virtualization captures and creates a realistic, highly scalable and responsive version of virtually any system the Performance Lab needs to connect to, using whatever tools they currently use for load testing - whether that is HP Performance Center (LoadRunner), IBM Rational platforms, Micro Focus' Segue tools, or some of our own system-wide or Web 2.0 test & validation apps.
These virtual services become assets that look and act exactly like the real thing - but they are very lightweight, low cost simulations, with very stable data scenarios, adjustable or variable response times, and the ability to be ready for reuse 24/7 at a moment's notice. The resulting virtual performance lab allows these software to be continuously performance tested at a component level, ensuring that when the pieces come together, they exceed SLA expectations.
Let's define a goal together: No more constraints on capacity in the performance lab. Hope you join us June 24 and find out how to get there.
Comments