Workload Placement Concerns in a Hybrid Environment

By Len Rosenthal, CMO


I just finished reading an interesting new Gartner study on developing a practical workload placement strategy.  Gartner lists 10 business factors when rationalizing where workloads belong.  Some are obvious, and some, like regulatory issues, are less so.

Of the top 10, four are directly related in some fashion to performance and availability issues, and one is closely related.  Here are those five issues, and how VI can help …

  1. Performance – impact on poor or variable performance? VI can help by identifying bursty applications which by their nature are hard to predict, and often caused by the variability of your customer’s use.
  2. Latency – Gartner breaks this key performance indicator out separately. VI reports on latency at a very granular basis.  If you’re currently getting latency reports based on 5-minute polling averages, you are almost certainly missing application / user problems.  And moving to the cloud won’t help.
  3. Service Continuity – is the perception of 100% availability required? Our perceptions of continuity have changed.  A seemingly minor bottleneck can cause a slowdown that customers will perceive as an outage.  Everyone promises real time alerts; talk with us about real real-time to see the difference between “5-minute” real-time and sub-millisecond real time.
  4. Availability – Gartner breaks this out from Service Continuity as more inward focused, the actual impact on business ops. Like with Service Continuity, by seeing I/O at the physical layer, and correlating with the traditional metrics you get from software monitors, our customers can head off problems before they bubble up to application slowdowns.
  5. Reputation – what’s the impact of outages or slowdowns? I agree with Gartner; it merits separate consideration.  You almost certainly have applications which can withstand performance problems, and applications which can’t.  We can show you examples of our customers who properly monitor their infrastructures and who quite literally eliminated sev 1 infrastructure-related problems.  If this sounds like a bold statement, and you’d like to see what they really do, email us and I’ll put you in touch with the right folks to talk to.