The Riverbed Blog (testing)

A blog in search of a tagline

Archive for June, 2009

Eating our own dogfood, part 2

Posted by riverbedtest on June 24, 2009

In my previous blog, I raised a few questions about Cisco's internal use of their WAAS product.  While I never doubted that Cisco has some WAAS devices deployed in their internal network, I did ask important questions about whether internally, Cisco is truly using and embracing WAAS to the same extent that they want their largest customers to.  Larry Chaffin echoed my questions in his Network World blog:

http://www.networkworld.com/community/node/42889

Subsequent to the discussions on Larry's blog, Feng Meng provided a few vague pieces of information about Cisco's internal use of WAAS in his Cisco blog:

http://blogs.cisco.com/datacenter/comments/wan_optimization_and_application_acceleration_case_study_-_cisco/

From the few tidbits of information Feng has offered, there are few things we can now conclude about WAAS:

1)  While Cisco aggressively encourages their customers to deploy WAAS to hundreds of sites, Cisco internally has yet to deploy it internally at such scale.

2)  While Cisco is a US-based company, they have yet to deploy WAAS in their domestic locations where the majority of their employees are located.

3)  The scale and scope of Cisco's WAAS pilot remains unclear, as Feng gave no details as to how many employees actually used WAAS in the pilot.  Caution is in order, as conveniently left-out details can be used to lead readers to inaccurate conclusions.

4)  Since Feng complained that Cisco has almost 10,000 servers, it seems Cisco is burdened with the same "server-sprawl" issues that are typical of other large enterprises.  But WAAS apparently has not allowed Cisco to consolidate servers and eliminate data centers as Riverbed's similarly-large enterprise customers have been able to do.  Indeed, the number of servers in Cisco's IT infrastructure appears not to have changed from the 9360 that existed before WAAS, as documented in Cisco's WAAS case study from June 2008.

As Feng lamented, I can understand how WAAS has been prioritized below other projects such as Telepresence by the internal Cisco IT team.  As a former IT manager myself, I certainly would have done the same thing.  But IT organizations that do not report to Cisco have the option to consider Riverbed.  Unlike WAAS, Riverbed has been deployed to the largest IP networks in the world–including networks that are much larger than Cisco's internal network.  And also unlike WAAS, the Riverbed solution has a track record of consistently delivering positive ROI–often within a period of a few months–for networks of all sizes.  Many IT organizations would readily place a project like that at the highest priority.

Posted in Uncategorized | 1 Comment »

SharePoint – Made for the WAN?

Posted by riverbedtest on June 18, 2009

These days, about one in three companies that I talk to use Microsoft SharePoint in at least one part of their organization. Some have made it the central source of information for their business; others are still exploring how to effectively use it. Regardless of the size of the deployment, however, it seems that performance is consistently a problem. Why is that?

SharePoint, unlike legacy Microsoft applications, was specifically written with the WAN and the Internet in mind. (In fact, if you head over to Microsoft TechNet, you can find an article called "Optimizing Office SharePoint Server for the WAN.")

Microsoft makes a nod to WAN optimization devices in general, but doesn't really talk about the extended benefits of WAN optimization. I think that ESG does a nice job of combining some of their own survey results on SharePoint with some impact analysis in their recent paper, "Why SharePoint Needs WAN Optimization Solutions."

The crux of ESG's paper is that currently SharePoint deployments are proliferating in a somewhat uncontrolled manner throughout organizations – whenever a site wants to use SharePoint, they deploy a local instance. In fact, they report that about 75% of organizations surveyed have deployed SharePoint in more than one physical location!

Instead, according to ESG, organizations should consider running it all centrally and using WAN optimization to provide the needed performance to distributed users everywhere. The end result is potentially significant cost savings by avoiding distributed systems and management. Again, based on their survey results, a majority of respondents said that with WAN optimization they experienced improved response times and more reliable data access.

We've seen organizations like Avanade, who are specialists in SharePoint, adopt WAN optimization as a key element for SharePoint, and I believe it will become more prevalent for organizations to design SharePoint systems with WAN optimization in mind. What about your organization? Do you think WAN optimization would significantly change the way you use SharePoint? Let me know your thoughts.

Posted in Application Acceleration | 3 Comments »

Eating our own dogfood

Posted by riverbedtest on June 16, 2009

Alpo In 1988 an email was circulated internally within Microsoft about the importance of “Eating our own Dogfood.”  Usage of the phrase spread not only within Microsoft, but throughout the high-tech industry, where the phrase is commonly used to communicate an important principle.  The principle basically states that any vendor who expects customers to use their products should also be using these products internally, in their own internal IT infrastructure.

 

As with most other companies in the high-tech industry, Riverbed practices the principle of “Eating our own Dogfood.”  With about 1000 employees, Riverbed is a medium-sized company with about 30 office locations around the world.  Every one of those locations have a Steelhead appliance for optimizing WAN connectivity, and every Riverbed employee has a Steelhead mobile software client installed and in use on their laptop or workstation.

 

However, when it comes to WAAS, Cisco seems to believe the dogfood principle does not apply.  Cisco touts WAAS as a scalable enterprise solution that can be deployed to hundreds of sites.  But they have so far failed to show that they have deployed it at such scale in their own internal enterprise.  Although Cisco believes WAAS is good for their customers, apparently they've been unable to demonstrate how WAAS provides value for their own internal employees.

 

Any customer considering a significant investment in WAAS should ask Cisco these questions.  How many WAAS devices have been deployed in Cisco's own internal IT network?  Does Cisco have their own operational WAAS deployment that is at least as large as the one that they would propose for your network?

Posted in Uncategorized | Tagged: | 9 Comments »