The Performance Zone

MPLScon in New York

A colleague of mine was at MPLScon in New York City last week and was commenting on how well attended it was – both by telecommunications providers as well as enterprise corporations.  There are many challenges and opportunities in deploying MPLS and this is a premier conference to help learn and flesh out things that need to be taken into consideration before, during and after enterprises roll out MPLS as their choice for WAN communications. 

For telecommunications providers, they had sessions on “Traffic Engineering in a Tier 1 Network”, “Using MPLS to Optimize and Mix Residential Triple Play and Business Services Networks”, and “The Role of Policy Enabled Subscriber Management.”  For enterprise attendees, there were sessions focused on “Enterprise Buyer’s Guide to MPLS-VPNs” and “The Enterprise Routing Guide to MPLS VPN Service Migration.”   There were also some joint sessions offered that covered a number of different Case Studies, MPLS Trends, Convergence and MPLS-VPNs – a little something for everyone.

My colleague offered a session in the enterprise track on “Managing Network and Application Performance through an MPLS Cloud.”  I was very interested to hear the informal responses (by show of hands) to some general questions.  After many years of hype, it appears enterprises are starting to role out MPLS here in North America.  But it seems to be still in the early stages, with most companies supporting some MPLS enabled regional offices with most other remote offices still using legacy Frame Relay and ATM.   It also seems many enterprises are incited to migrate to MPLS as part of their VoIP roll outs.

Industry analyst, Dr. Jim Metzler discussed the direction of MPLS in a recent IT Impact brief called The Movement to Deploy MPLS and sites a recent survey of 280 enterprise IT professionals conducted by Ashton Metzler and Associates on the evolution of MPLS in enterprise networks. 

Over the next few weeks, I will cover a number of topics and trends related to MPLS.  I would love to hear from you if you are in the middle of an MPLS roll out – how is the cut over progressing, what issues are you finding, what should other enterprises plan for and look out for in their conversions?   

Performance management and virtualization comments

One of the reasons that I love my job is that I get the opportunities to interact with a lot of customers and prospects. This week we held a couple of regional user group meetings and got some very insightful feedback from our customers. Two observations particularly stand out.

One  customer said that they found the nGenius Solution helps them significantly in managing the performance of their virtualized network. This comment is not only a compliment about what we are doing but also a validation of  the need to manage the performance of  interconnected virtualized resources built on the physical infrastructure. That is the cornerstone of our Application Fabric Performance Management strategy.

Another customer asked a related question: How does NetScout see its role in relation with tools dealing with element health and network management? The answer again is AFPM. With this new architecture, we are able to monitor the performance of applications in the context of their underlining fabric resources, be it server, network, storage, or middleware.

It is really exciting to see how performance management is evolving in this virtualization age. What are your thoughts on this topic?

Customer Web Service based App and Virtualization story...

Recently, one of our product managers was visiting with a customer that had experienced a situation that is likely to become more prevalent as web services and virtualization become a cornerstone of IT architectures.

Our customer had been working on the release of a new application. In advance of its wide-scale deployment, the application had undergone testing in the lab and demonstrated itself to behave in a manner that would not have a negative impact on the network. While we’ve all seen applications that behave in quite a different fashion when they are deployed on the WAN as opposed to the clean LAN-based test environment – that’s only a part of the story.

The application was web-based, and utilized video to provide training to their employees. Right about now, you’re probably thinking that video was the problem. You would only be partially correct! The real challenge came in the fact that there was a corporate directive that employees must use the application to expand their product knowledge and improve their operational efficiencies.

While testing the application in the lab revealed that the application would have some impact on the network. The sheer number of users combined with the large amount of video services available behind the application drove utilization well beyond acceptable limits. This behavior began to have a negative impact on other business applications. Adding fuel to this fire was the fact that the application was web-based and was initially seen as an increase in http traffic. However, the nGenius system administrator was able to quickly create a URL-based application definition and extract this application from all other web-based services. Armed with this information, IT was able to demonstrate the impact and usage of the new application to management and adjust QOS levels accordingly.

As SOA implementations begin to be rolled out over web-based infrastructure, we are likely to see more of these scenarios appear on our networks. Virtualization provides application developers with much greater flexibility and capability, but the cost can be significant in terms of network impact. What’s worse, if we aren’t in a position to be able to clearly identify the applications contributing to the increases in network traffic, the network will once again be the primary point of blame for poor application performance.

High Definition Video Conferencing and High Definition Performance Management

As a technology professional and a multi-media buff, I get to see a lot of "cool" video tools and toys, many of which have fallen short of their promise. With that mildly jaded view, one 'cool thing' coming out of the TelecomNEXT show I attended a few weeks ago was telepresence.

Cisco revealed plans for a telepresence enterprise video conferencing system, which aims to be more “lifelike” with life-size high-definition video and directional sound technology. This sounds similar to Hewlett-Packard's new HP Halo Collaboration Studio offering. Just imagine how cool it would be to meet colleagues, partners, or customers on the other side of the Pacific as if they were face-to-face with you. I would not mind joining the “infrequent flyers” club at all. Sound far-off? According to Cisco, it could be available this summer or fall.

While the deployment of bandwidth-rich IP video and voice applications such as “telepresence” have the potential to improve productivity and operational costs, it will require a more robust network infrastructure. More enterprises will probably need to move to a higher speed network infrastructure, 10 Gigabit in the core, in order to meet greater demand for network-bandwidth consumption of those multiple-megabit IP video streams. I know - what a shocker - Cisco is pushing companies to upgrade their networks.

The other side of the coin though is management. Enterprises will definitely require sophisticated application and network performance management tools to assure their business services are working well alongside with this pseudo-real-life video application running across the network. Even if a company has a huge 10 Gigabit pipe running between major corporate sites or data centers, bandwidth might not be a problem yet, but monitoring what other applications are running across that pipe will still be important. On the LAN side of the stream, network professionals will surely need to see more granular performance information to assure the smooth delivery of high definition video streams. In other words, high definition video conferencing needs high definition network and application performance management.

Closing out the show and a fiscal year

It has been a busy week here at Interop and I have enjoyed talking with so many different networking professionals!  Clearly, the message of the show was infrastructure change to adapt to the growing complexity of applications throughout the network.  Web based, n-tier, and SOA applications that deliver business services to end users is definitely going to change the way IT organizations manage performance across the network and application fabric. 

Many folks have stopped by our booth (#2249), test driven our nGenius Performance Manager and our new nGenius Analytics, and it is great to hear their enthusiastic feedback.  If you missed the show, be sure to come to www.netscout.com to learn more about our direction.

I also took a moment to reflect on our last year.  When we announced our strategy yesterday for AFPM and demonstrated how we are going to help IT organizations really effect a change in their current MTTR for problems that affect end users in their networks, it was the culmination of a major R&D effort for NetScout over this past year.  And it will be exciting to play a part in these next 12 months delivering technology that automates alarm and threshold setting, that automatically identifies and diagnoses performance problems, and that helps you implement corrective actions before your users are impacted by degradations.

Well, off to catch another plane – it will also be good to be sleeping in my own bed soon, too!  Hard to get much sleep here in Las Vegas – but that's all I will say – because as you know – “What happens in Vegas, stays in Vegas!”

nGenius Analytics announced at Interop Vegas

Yesterday was a lot of fun!  It has been a while since we had a booth at Interop as the show lost its focus over the years. The show is much more focused on networking! It was refreshing to see the number of high quality folks visitiing our booth and not just looking to get a t-shirt :)

Today we announced: our new nGenius Analytics products; a customer, CN Railways, who has been a pilot user of it; and our direction for Application Fabric Performance Management (AFPM).  Please be sure to come by our booth (#2249) at Interop in Las Vegas to hear more about all these initiatives. 

NetScout unveils new look at Interop Vegas

Interop – Day 1 – NetScout introduces a New Look!!

Please be sure to come by our booth (#2249) today at Interop in Las Vegas. We are proud to unveil our new look and company logo! It is sure to be eye catching!

While you are there, stop long enough to learn how to overcome the challenges to monitoring MPLS networks. I will be there today as well so stop in an say hello!!

Stay tuned for moe info on Interop Vegas!

p.s.
NetScout will also unveil it's new web site tomorrow @ 8:30 EST, so check it out at @ www.netscout.com.