Countless science fiction books and movies, from “The Matrix” to “1984” to “2001: A Space Odyssey” feature computers and networks that are completely omniscient with the power to see and hear everything users do.
While today’s networks may not talk back like the HAL 9000 or be run by sinister alien robots that want to use humans as a power source, next-generation networks today have the ability to understand the applications being run, see the end users of those applications, and even deduce what devices they are using.
Innovations in next-generation networks are driving new sets of capabilities in three major areas: increasing visibility, prioritizing traffic, and optimizing performance. Cisco provides these three things combined in the AppVelocity Network Service. A ‘good enough’ network can’t deliver these capabilities.
Is Big Brother Watching Your Network?
Isolating network from application performance issues is a classic challenge in IT. Big brother may not be watching, but if your ERP system is performing poorly, how can you tell if it’s an application issue or other traffic on the network causing this performance problem?
With a good-enough network, you can only roll out sniffers and probes and dig into the network traffic. But if you are troubleshooting a performance problem from the United States and the problem is in a branch office in Korea, it may be very difficult to get the necessary tools deployed in that branch office.
A better answer is to provide integrated tools that allow for the capture and analysis of traffic to help detect and isolate problems. Cisco does this with a family of technologies that are all managed under the Cisco Prime Management umbrella.
The first of these, the Network Analysis Module (NAM), is embedded hardware and software that can perform granular traffic analysis, rich application performance metrics, comprehensive voice analytics, and deep insightful packet captures to deliver a comprehensive performance assurance solution.
In addition to NAM, Cisco router and switch products provide Network Based Application Recognition (NBAR) support. This allows the network device to detect traffic types so you can create strategies to prioritize traffic based on the application type.
Network Traffic Jams
Vendors of Good Enough Networks will saythey can prioritize traffic with basic Quality of Service (QoS)QoS. While QoS isa important element, it is only part of the solution. There are two other critical elements that are only provided by an Enterprise Next Generation Network. The first of these is NBAR mentioned above. NBAR provides built-in application intelligence in routers and switches that allows for the detection of application types on the network so the network administrator can build simplified rule sets to determine which traffic should be prioritized by QoS. Without NBAR, the typical network administrator is required to become a protocol by protocol expert at a bits and bytes level to create the right prioritization strategy. That approach doesn’t scale.
The second critical element is the ability to prioritize traffic when things aren’t “business as usual.” While QoS is good, it only provides prioritization as traffic is leaving a network element and that doesn’t help when there are problems deeper in the network that are causing the application performance problem. This is particularly an issue in branch offices that are connected to the headquarters with multiple connections (i.e. multiple MPLS providers or one provider with broadband VPN backup).
To help solve network performance in a reactive fashion, Cisco has built a features called Performance Routing (Pfr) which can detect issues such as high latency, jitter, packet loss, and make sophisticated decisions to help route around performance problems allowing critical applications and voice/video services from being severely impacted. It is the marriage of visibility and prioritization that provide a unique set of reliability enhancements for the enterprise next-generation network. How does a good-enough network deal with this? It doesn’t!
But what do I do when I just don’t have enough bandwidth? Well that leads us to optimizing performance ….
Optimizing Network Performance
When bandwidth is an issue, the typical answer has been to throw more bandwidth at it. This would be the good-enough network approach. But some bandwidth (i.e. WAN) is expensive, so throwing bandwidth at the problem is a poor answer at best. Advanced compression services and protocol optimization that optimize bandwidth and accelerates performance is the better answer. Cisco provides this through its family of Wide Area Application Services (WAAS) which can be embedded into the ISR G2 family of access routers. For those customers only needing compression capabilities the G2 can offer this through WAAS Express, a software only option. For more sophisticated protocol optimization and compression capabilities, the G2 offers WAAS on the SRE hardware module.
With the WAAS solutions, network traffic is classified and optimized with a sophisticated set of compression services and protocol optimizers. The result is increased bandwidth availability to branches and enhanced end-user experience and productivity. So it pays for itself in a very short period of time.
Optimizing performance is not just about solutions in the branch, how do you bring the technology into the Data Centers where the applications reside? This is another advantage of an Enterprise Next Generation Network in that it looks at how services can be delivered in new and creative ways. With WAAS we have a great example of this delivering Data Center WAAS services through vWAAS as a service running on a virtual machine in the Data Center. This puts acceleration services closer to application and simplifies the deployment model.
It may not be big brother or HAL watching, but an enterprise next-generation network knows more about the traffic on the network and is able to optimize the analysis, troubleshooting, prioritization, and performance of applications on the network. All of these are factors that are critical to the business value of the network and require significant vendor innovations embedded into the network. A good-enough network is only going to leave you blind to what is happening on your network and doesn’t provide tools to solve real performance, endpoint, and application challenges. Another example of ‘good enough’ not really being good enough.
Seven myths down…and we’re done. For those of you who have been following this series, I hope it has given yourself something to think about in how you approach the architecture of the network. While the good-enough vendors sees the network as a utility for basic transport, we continue to drive the R&D investment in a network-generation enterprise network, creating a network to respond to the needs of business both today and in the future.
Note: The seven myths are outlined in a recent white paper from Cisco:Debunking the Myth of the Good Enough Network.
Mike has been with Cisco for 14 years and is currently responsible for working with customers and partners. In his current role, he helps shape the future direction of the network and how it delivers business value to customers and partners. Prior to this, Mike was the Vice President of WW Enterprise Technical Sales Strategy where he was responsible for developing technical and competitive sales strategies in support of Cisco’s Enterprise go-to-market strategy. Mike has held a variety of other positions at Cisco, starting as a Systems Engineer where he supported Cisco’s entrance into the switching market.
Latest posts by Michael Rau (see all)
- Application and End Point Ignorant: Myth# 7 of the Good-Enough Network - June 21, 2011
- Myth # 6 of the Good-Enough Network: Acquisition Cost - June 14, 2011
- Myth # 5 of the Good-Enough Network: “Basic Warranty” Myth - June 7, 2011