The Riverbed Blog (testing)

A blog in search of a tagline

Is Cisco still trying to eat their own dogfood? I thought they had given up…

Posted by riverbedtest on January 28, 2010

Cisco first began selling WAN optimization in 2004.  Over the 5+ years that they have been aggressively promoting and selling the WAFS/WAAS product to their loyal customers, Cisco has been unwilling or unable to use WAAS themselves, in their own network.

The discrepancy between Cisco's words and actions must have been uncomfortable, because in 2008 Cisco finally made formal plans to deploy WAAS in their own network infrastructure.  These plans were documented in a WAAS progress report posted on Cisco's website in 2008.  An excerpt from page 6 of this document states the following:

Cisco has approximately 300 global offices that have less than 45 Mbps bandwidth.  Cisco partner IBM will deploy Cisco WAAS in these offices during the remainder of fiscal year 2008 and part of fiscal year 2009.

2008 and 2009 passed, and there was still no sign of an operational WAAS deployment in Cisco's own network.  I wrote my first dogfood blog in June 2009, wondering how they could sell WAAS without using it themselves.  Larry Chaffin similarly asked in his Network World blog ("Could the Rumors be true about Cisco?  They don't use WAAS in their own network?").  No response from Cisco, except for a few muddled references to "pilot testing" in some of their branch office sites.  Cisco even refused to comment on the progress of their WAAS deployment when confronted at their Cisco Live! conference.  The internal WAAS deployment effort apparently was not going according to plan.

At this point, I thought Cisco had given up on their WAAS deployment, just as many of Cisco's former WAAS customers have.  But unlike former WAAS customers, Cisco's IT department is probably not allowed to turn to Riverbed.  Then I saw a new blog, dated 25 January 2010, on Cisco's website:

The internal Cisco WAAS implementation continues to  expand broadly . So far IT has deployed WAAS to 200 Cisco remote offices worldwide.  When fully deployed over the next six months, we will have implemented WAAS in approximately 300 Cisco offices and eight data centers worldwide.

Here we go again.  Years after Cisco's previous aborted attempts to deploy WAAS, it seems Cisco is going to make another concerted effort to deploy it.  But Cisco's blog also raises even more unanswered questions, such as:

1)  The blog states that 200 WAAS devices have been deployed in Cisco's branch offices, but none seem to be deployed in Cisco's data centers.  What good is WAAS if they aren't optimizing network traffic to the data centers, where the majority of their application servers are located?  (and of course, the data centers are where the WAAS per-peer data store experiences severe scaling problems…) Are these "deployed" WAAS devices doing little other than optimizing a small amount of regional traffic?

2)  If Cisco's branch offices have less than 45Mbps of WAN bandwidth as documented in Cisco's 2008 WAAS progress report, then why are they primarily deploying the WAE-674 to these offices?  The WAE-674 is rated by Cisco for 90Mbps of WAN bandwidth, so why isn't Cisco using more of the smaller and more cost-effective WAVE models?  Does Cisco lack confidence in the performance specs published in their own WAAS data sheets?  Is Cisco internally aware of a WAAS performance or scaling issue that isn't reflected in their product documentation?

3)  Cisco's blog boasts about their hopes of saving $850,000 over 3 years from their 300-site WAAS deployment, as a result of consolidating servers and storage.  $850K over 3 years???  That's hardly impressive when you consider how Riverbed customer Lantmannen expects to save $60 million over five years after deploying the Steelhead solution to 300 sites.  Furthermore, note that a single WAE-674 appliance with the WoW virtual blade enabled costs $28,500 at list price.  300 of these deployed to Cisco's branch offices amount to $8.55 million.  Even at a 50% discount, Cisco's branch deployment would amount to spending $4.275 million dollars in order to save $850,000 over three years.  After adding a few additional million dollars to the bill for the data center WAAS devices, you end up with a hugely negative ROI.

Clearly, Cisco is not attempting to deploy WAAS for the ROI, which is hugely negative (even after Cisco's 85% discount for internal transfer pricing), but to convince their concerned customers that they can eat their own dogfood.  We wish Cisco luck on their renewed WAAS deployment efforts; perhaps they will get farther than their previous attempts. But of course, if this upcoming deployment effort doesn't work either, they can always write another new blog about planning to deploy WAAS in 2012…

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: