Virtualization in LIMS

Jason had the following comment about my Virtualization in LIMS blog.

Q: One hurdle that I can think of is the ability to validate the VMWare software. What validation approach would you envision for snapshot management?

A: I don't want to position myself as a validation expert, but here is my opinion. I do not believe that virtualization software would require a different strategy than would be normally used for validation.

Validation is performed on a processes- not on software itself. You would not be able to validate VMWare or any virtualization software, just like you can't validate Microsoft Word or Microsoft Excel. Instead, you can validate the process of using the LIMS (which would include the virtualization software).

So, to deal with specific functions of the virtualization software, you would have to:

1. Determine specifically how you would use the virtualization functionality in the context of using the LIMS software
2. Test that the software functionality performs as expected (according to specifications) in the OQ/PQ
3. Use the functionality within the bounds of testing.

In my opinion, this is the same strategy that you would use for any software or hardware component that is part of a LIMS.

Further, I believe that using virtualization software or hardware may actually make some elements of maintaining the validated state of the system easier, albeit at the expense of a little extra work up front. I believe this is possible because the use of virtualization software can obscure the specific hardware that is used in the LIMS system. Normally, you would expect that a change in the hardware used (replacing a motherboard or CPU, for example) would require re-qualification of the new hardware prior to use. If it can be established and qualified that the virtualization software works across different physical hardware, it is feasible that the LIMS would stay in a validated state in the midst of component changes.

I think the same can apply to the snapshot functionality you reference.

Zen and the Art of LIMS Troubleshooting

Joel Limbardo, who always has interesting commentary on the LIMS community, provided the following comment on my Zen and the Art of LIMS Troubleshooting Blog.

Q: Odd, I know individuals who have advanced degrees in chemistry that utilize the scientific method daily but cannot operate their computer for anything beyond simple e-mail. This article is nice, but did you consider that each and every support call costs money? The objective of a software engineer is to create software that produces virtually no need to call any type of support technician. I agree that a balanced, strategic approach to problem solving will aid the support professional, but the answer really is not the pure 'scientific method'. Instead, it is a combination of building robust software and running through a great many failure scenarios before that product hits production. The result of those failure dry-runs will be troubleshooting scripts that the support professional SHOULD USE to find and diagnose the problem according to the development specifications. Ask yourself this -- should your Xerox support tech. be using the scientific method? I thought not.

A: I agree with a lot of Joel's comment. Good engineering techniques should be used to reduce software defects, which reduces the load on a support desk. Further, dedicated support desks should be equipped with a knowledge base that allows them to resolve common issues quickly. I would even want that knowledge base to be available (at least partially) to customers so that they can solve some of their own issues. If a knowledge base or script reduces support costs, allowing customers to resolve some issues themselves saves even more money.

It is a mistake, however, to equate "support" and "troubleshooting" with "defects". Sure, diagnosing and repairing software defects is a big part of a Support desk, but it is only one part. Providing good support for a customer means helping the customer use their software or hardware effectively. Some companies regard Support desks as a way of building rapport with and loyalty from customers. In other words, instead of a group that focuses on problems, a Support group can be a group that focuses on customer support.

Regarding the Xerox support tech. that is servicing my equipment, allow me to ask Joel's question a different way: Should your Xerox tech. be able formulate a hypothesis, isolate variables, produce experiments, and draw valid conclusions about the way things are working or should they simply be only capable of following previously created troubleshooting scenarios without factoring in my unique situation? Well, the answer depends on what I want from the service provider. Do I want them to reduce their service costs to as close to zero as possible or do I want them to build a relationship with me so that they can help my business succeed.

My experience is that many LIMS vendors want the former (but say the latter) and most LIMS customers want (and pay for in service fees!) the later.