Cisco VP lets us know WAAS up!

George Kurian, VP of Cisco's application delivery division, recently held a conference call where he provided a state of the WAN optimization market along with an overview of Cisco's WAAS 4.1 release.  While I applaud George for coming public to address key issues involving Cisco's WAN optimization product, there were many misleading and inaccurate statements.

For starters, George was asked how he would like to respond to Gartner's recent report that tells customers to proceed with caution with regards to Cisco's WAAS product.  George responded with 3 key rebuttal points:

-  The report was based on code 4.0 and some the issues have been addressed in the 4.1 code
-  The sample was very biased as it only included about 3% of their customer base
-  Issues like asymmetric route handling are not specific to WAAS and occur with all vendors

Personally I don't believe George when he speaks for his customers, especially when they are telling us that they are still having stability problems with 4.1 --  how about we both agree to let them speak for themselves? Maybe in a face to face bake off between our products?

Riverbed continues to win 9 out of 10 head-to-head bake-offs against Cisco and the stability issues are a major component of the selection criteria.  Cisco can continue working on addressing stability issues, but at some point, one has to look at the fact that the underlying architecture and foundation of the product is not suited for large deployments.  Check out my blog post about architecture considerations.

On the 2nd point, a 3% sample out of a reported customer base of "over 3000" is pretty significant.  Asking 100 customers about their experience with Cisco WAAS and getting overwhelmingly negative feedback speaks volumes.

On the 3rd point, George has not been doing his homework.  If he had, we would know that Riverbed does indeed support asymmetric routing environments, a unique advantage over other offerings including WAAS.  The Steelhead appliance will detect and report when an asymmetric routed environment is present and we have features like connection forwarding, which allow a set of Steelhead appliances to share TCP session information so they both understand the session flow characteristics and can work even in the face of asymmetric routing.  This is one of the many advantages Riverbed has and is one of the reasons why our product is able to be deployed in even the most complex networking environments.

That's enough about the Gartner rebuttal.  Other points made needing refutation include:

Cisco WAAS deployment success?
- Dozens of customers deployed at more than 75 locations
- More than 10 customers deployed at over 500 locations
- 55 of the fortune 100 have deployed WAAS

You might like to know who these customers are!  If they are so successful, why is it that a majority of these customers not referenceable?  In contrast, Riverbed has more than 200 customers that we can provide to our prospects as a reference.  Riverbed has both the largest scale and most number of referenceable deployments, with many of them not only talking about their positive experience with Riverbed, but also about their poor experience with evaluating Cisco WAAS.

Cisco is the WAN Optimization market leader?
- 2Q 08 Garter report states market leading 23% share for Cisco WAAS

This is where the details start to get a bit confusing.  There are really two pertinent market categories covered by Gartner.  One category is the Wide-area Optimization Controllers (WOC) and the other is the Advanced WOC category.  Riverbed's product is properly positioned in the advanced WOC category where we have a lead over Cisco in this market.  The important point here is that collecting market data is sometimes achieved by asking vendors what their revenue is.  For Riverbed, WAN optimization is all we do and we are a public company so our numbers are clear.  For Cisco WAAS, they can tell Gartner any number they want and Gartner has no way of verifying.  Projecting your market share leadership without validation of your specific revenue for this market is misleading.

When asked, why do you win against Riverbed, George?
George responds with that their solution is more comprehensive.  He goes on to say that it is better integrated with the network architecture and works better with services like QoS. 

This is the same old argument that Cisco has had all along.  You should choose Cisco because your network is Cisco right?  The fact is that Riverbed integrates very well with Cisco networks.  Just ask our 4500+ customers in what types of network environments (with QoS, VoIP, etc.) do they have Riverbed Steelhead appliances successfully deployed? 99% of our customer base uses Cisco as the networking infrastructure and our stuff just works.  This argument is easily diffused by our customers references that have Cisco networks.

WAAS Mobile is seeing good adoption with almost triple digit customers  
Almost 100 customers?  Contrast that to Riverbed where our groundbreaking Steelhead Mobile product has resulted in more than 15% of our customers adopting this product in less than 13 months of its existence.

Our mobile product was built from the existing Steelhead appliance code-base that has been successfully deployed at thousands of customer environments for years.  Steelhead Mobile leverages the same Steelhead appliance data center infrastructure, the same layer-7 latency optimization, the same data deduplication algorithms, and the same administrative capabilities like auto-discovery.  The result is a mobile acceleration platform that extends Riverbed's superior performance for the data center and branch office to the mobile worker.  Contrast that to Cisco's mobile product, which is essentially a licensed version of ICT's Accelenet product that has zero WAAS integration and leverage and even requires a separate server platform at the data center.

There were many additional tidbits of mis-information during this short call, but the point has been made.  At the end of the day, some people at Cisco seem to have Riverbed in their cross hairs and I am looking forward to the challenge.  What we have is David vs. Goliath all over again.  However this time, while Goliath's size is huge, David has the better weapons and the army of happy customers to help fight the battle.

Comments

[Asymmetric_route]

Connection forwarding is in and of itself not a one-trick-pony for asymmetric routing. There are certainly scenarios where connection forwarding falls apart or is otherwise not the most optimal solution, and other vendors including Cisco have features as well as design best practices that support asymmetric routing equally well. I believe you may have taken his statement out of context.

[Mead]

As a Riverbed customer who tested Bluecoat, Cisco, Citrix, F5, and looked at a number of different products I'll say our early deployment has had its bug issues with Riverbed but the product is far easier to deploy and works far better than the WAAS product or any of the other vendors products as we tested them.

I really wonder if Cisco is tooting their horn just because of the negative report from Gartner as well as the numbers of companies who struggle with their WAAS deployments. I hear rumors from fellow tech friends where some of there deployments are deals where its tossed in with a large router purchase for free.

I hold concerns if thats true and they tie that to market share or sales as thats not very realistic.

Also tying the WAAS and the Router isn't a very easy thing to setup or manage. God forbid if you do Policy based routing based on QOS and try to toss in WCCP. Don't even ask about inline abilites of there products and asymetric routing support. Finally the last nail in the coffin would be the limited WCCP support in the IOS line. There is a lot of complex issues they need to fix and simplify before I would say their product is ready for enterprise use.

That's from a customer perspective.

Not to bash Cisco, I love their routing and switch products but there are some technologies they just haven't been able to seriously keep up with or intergerate correctly after they bought them.