Journey through Enterprise IT Services
In Journey through Enterprise IT Services, Nadhan, HP Distinguished Technologist, explores the IT Services industry, and discusses technology trends in simplified terms.

Do you have the right test environment for cloud computing?

Recently I blogged about enrolling my son into college. I had been to the Open House conducted by the Mechanical Science and Engineering department of the University of Illinois at Urbana Champaign where Professor Emeritus James Phillips was kind enough to give us a quick overview of the testing facility at the Talbot Lab. This is just one of three such facilities across the United States. The Southwark-Emery testing machine will load specimens in either tension or compression to 3,000,000 lb (13 MN). Research Laboratories like these provide access to state-of-the-art facilities for tension and compression testing of large components and structures.

 

Prior to the real-life deployment of such structures, it is vital to ensure that there are robust test environments like these so that we can be wiser about the loads that these structures can withstand. Software is no different. Software needs to be tested for load as well especially when the applications are deployed in cloud computing. What kind of environments do we have for testing applications in the cloud?

 

As Charles Babcock mentions in a recent Information Week article -- HP Offers Test Cloud packages, HP provides a low-priced test bed to help newcomers launch cloud workloads.

 

HP offers a helping hand at the outset and lets customers start experimenting with their workloads for up to 16 weeks. New customers are provided with established services, including Web servers and a database service while working with whatever the customer brings (including their own database systems) in alignment with the Your Cloud Your Way theme.

 

The customer's workload can run in a virtual machine on a server shared with other customers -- a multi-tenant environment -- which allows customers to validate real-life scenarios to address any underlying security and privacy concerns they may have.

 

When building solutions, it is always beneficial to understand the extent to which they can meet the functional requirements for the complete spectrum of transaction volumes -- low, average, high and peak. The consumer of these solutions really does not care about the transaction load within. Imagine yourself standing at an airline counter to be served by the agent at the counter. What matters to you is the fact that you are there waiting for service regardless of whether it is 2 customers ahead of you or 200.

 

It behooves the solution provider to ensure that the system functions at peak loads. However, this needs to be determined prior to the deployment of the solution. Waiting until such loads are realized in the production environment is too late. Who knows, if such testing had been done at the right time with the appropriate monitoring mechanisms in place, we may not have seen the sudden collapse of some concrete bridges in some metropolitan areas.

 

It is comforting, to say the least, that environments that can effectively simulate such production workloads exist -- whether it be the:

How about you? How do you test your application environments? Have you seen similar environments to the one at UIUC? Please let me know.

 

This post is part of the Knowledge Matters cloud series. Be on the lookout for our up and coming blogs on cloud and mobility. And you can always check out our previous Knowledge Matters articles on the HP Applications Services blog

 

Connect with Nadhan on: Twitter, Facebook, Linkedin and Journey Blog

Comments
tom.hall | ‎10-12-2012 02:59 PM

Nadhan,  interesting post and serves as another example on how our personal and professional lives intersect.  One statement got me thinking “The customer's workload can run in a virtual machine on a server shared with other customers -- a multi-tenant [testing] environment…”

 

If I am performance testing an application via a SaaS model,  and there are other clients doing the same (multi-tenant), how would I know if performance degradation is due to a problem in the code of my app or due to several other test workloads executed by others using the same physical machine? 

 

Thoughts?

Nadhan | ‎10-12-2012 03:25 PM

Great observation, Tom.  It is true that other workloads are likely to have some impact on the specific instance of the application being tested.  My follow-on thoughts:

 

  • As mentioned in the post, this is about the availability of an environment to experiment with the workloads.  So, there is a degree of trial-and-error essays that apply here.
  • The Test environment should have control mechanisms in place to determine the number of tenants supported at any given time.
  • Having this environment allows the performance testing of real-life scenarios where these applications are likely to be in a multi-tenant environment.
  • If it is predetermined that a multi-tenant environment is not conducive to a given application, I agree with you that the testing must be done in a similar environment without any other tenants to obtain more precise results from the performance testing exercise.

Connect with Nadhan on: Twitter, Facebook, Linkedin and Journey Blog.

NancyLichtle | ‎10-15-2012 01:13 AM

Nadhan, as I read your post, I clicked back to your Father’s day post , and it got me to thinking.  I’m wondering if you’ve captured your father’s insights electronically or on paper.  You point them out in terms of the enterprise (all of which support a cloud infrastructure!): Strategy, Context, Interoperability, Best Practices, and Frameworks – all excellent points. But I’d love to hear more about your father’s advice and your lessons learned. I think many would benefit from 1) your insights and 2) the ease of reading. And if it again ties back to our progress for the enterprise and cloud so much the better. You’re a remarkable writer and I look forward to reading more!

Nadhan | ‎10-15-2012 02:53 PM

Nancy, Thank you very much for your comment and kind words.  Since your question is related to the Father's Day post, I have responded to this comment on that post instead. 

 

Also, have you checked out my new Journey through Enterprise IT Services Blog?

 

Connect with Nadhan on: Twitter, Facebook, Linkedin and Journey Blog.

Leave a Comment

We encourage you to share your comments on this post. Comments are moderated and will be reviewed
and posted as promptly as possible during regular business hours

To ensure your comment is published, be sure to follow the community guidelines.

Be sure to enter a unique name. You can't reuse a name that's already in use.
Be sure to enter a unique email address. You can't reuse an email address that's already in use.
Type the characters you see in the picture above.Type the words you hear.
Search
Showing results for 
Search instead for 
Do you mean 
About the Author
Featured


Follow Us
The opinions expressed above are the personal opinions of the authors, not of HP. By using this site, you accept the Terms of Use and Rules of Participation.