It’s All About the Results!

We obviously like our product and are  pretty darn proud of it…all you have to do is browse this blog to see evidence of that.  But in the end, it doesn’t matter what I think (…my wife and kids will confirm this BTW) or my associates think or our partners think or even what Microsoft thinks.  It matters what our customers think…and while we proudly have many satisfied customers, few like to publically discuss what they’ve done and how they benefit from those efforts.  When they do we obviously want to climb to the highest virtual mountain tops (…this blog, our website, Twitter, etc.) and shout it at the top of our virtual lungs. So let the shouting begin…

We had the pleasure of recently adding Chesapeake Energy (CHK) to our customer list.  It started out with a customer that was using SharePoint pretty heavily, wanted to do more, but was hesitant because their SharePoint content databases were becoming increasingly difficult to manage.  They went to Microsoft for help and Microsoft in turn engaged us.  We participated in a Microsoft sponsored ADS at the Dallas MTC in March to discuss options and plan for a proof of concept.  The POC was scheduled for May, again at the Dallas MTC.  The purpose of the POC was to establish evidence around the following proof points:

• A 75% reduction in the size of the content database.
• No impact to OOB SharePoint functionality and end user experience.  Dozens of use case scenarios that exercised content within and outside of the SharePoint UI were executed.
• No measurable negative impact to storage and retrieval performance.  The assumption by CHK was that there would be some degradation in performance given the additional layer represented by StoragePoint.  There were no hard metrics established for this portion of the POC…they chose to go more off of “feel”.

CHK provided a image containing a copy of one of their production environments for use in the POC.  Once the POC environment was in place and running on MTC hardware we proceeded to install and configure StoragePoint.  That took all of 10 minutes which left the CHK team a little shocked.  They were expecting this would take the better part of the Day 1 morning.  With the install and configuration complete we proceeded to run the StoragePoint Externalize job to remove the existing content BLOBs from the content database and place them on the configured SAN.  Again, this process took considerably less time than was anticipated.  We then started the process to shrink the content database via SQL Server Management Studio.  We started that, ran it for a while, and then wrapped for the day because it was getting late.  When we arrived on Day 2 the database shrinking process had completed and we were ready to move onto use case and performance testing.  We sat around for a bit so the CHK team could digest the shrinkage results (…you’ll understand why below).  The first thing we did was kick off a full crawl.  They wanted to confirm that externalized content was still being crawled and what impact StoragePoint had on crawl/index performance.  For a point of reference it was taking them 6-7 hours to perform an incremental crawl on the content without StoragePoint (…they stopped doing full crawls because they were taking too long).  The full crawl finished and a few folks were skeptical it had worked…someone proclaimed “it only crawled the placeholders”.  Testing began and it was confirmed that the externalized content BLOBs had in fact been crawled.  A few more hours of use case testing were performed, everything from simply uploading a document from the UI to dragging a document from a Windows explorer window the desktop opening it, editing it, dropping it back into an explorer window and back into SharePoint thru WebDAV.  Every test they performed worked as expected.  So here are the results:

• 98.1% reduction in content database size, from 149GB to 2.6GB
• As noted above no impact to functionality or end user experience whatsoever.
• While no specific measurements were taken on the individually upload or retrieval of documents, the consensus of the CHK team was it was notably faster.  The one operation that was measured was crawl performance.  As noted above they were seeing a roughly 6-7 hour run time for incremental crawls.  It took 4 hours and 49 minutes to run a full crawl!  OK, take a moment to digest that and collect yourself.  It’s one of those your mileage may vary metrics, but we have yet to see a scenario where there has not been a mind-numbing improvement in performance.

The POC was completed in 2 days instead of 5 (...we strive to exceed expectations) and was deemed a success by all. 

CHK is in production with StoragePoint and looking forward to doing more with SharePoint as a result.  They will also be early consumers of our SharePoint 2010 version of StoragePoint.

I'm going to stop shouting now…my throat hurts!  I’ll let the customer step in and share their thoughts:

That quote is obviously a keeper…and was the inspiration behind our “SharePoint without Boundaries” tagline, as they really felt they had reached a boundary they weren't comfortable crossing.  StoragePoint removed that boundary in their mind.