Managing Application, VoIP and Network Performance in converged networks from edge to core

The Industry’s trend, to move from real client-server applications in favor of centralized, web-based programs is moving fast forward. In addition, many companies, no matter from which Industry, are in the process of moving the voice communication towards a partially or even fully integrated Voice over IP system. As a result additional data (VoIP control and call traffic) will be transported on the existing data network. Together with the fact that MPLS has reached many companies as well, there are new challenges IT Managers are confronted with today if they have aligned their IT strategies to follow these trends.

By Robin Gareiss – Executive Vice President, Nemertes Research

Think in solutions, not in problems…

Monitoring and troubleshooting converged environments also means that the different groups (“silos”) within the IT organization will have to grow a lot closer together because they are sharing the same infrastructure to deliver different services to the end users. As a result, there will be a larger overlap with respect to the monitoring and management tools the different IT groups (functions) will use. The fact that different management solutions will present similar information from different data sources in a completely different way does not help the communication between the members of the IT organization either…

Hence, there is a need for an integrated platform to manage Application, VoIP and Network Performance, which is capable of delivering broad enterprise or carrier network visibility as well as deep analysis and troubleshooting allowing IT organizations to maximize the value and performance of their IT infrastructure and deliver a superior end-user experience.

Challenges

Data Centers – Escalation Process Owner: There are multiple infrastructure pieces in a data center that can have a negative impact on the performance of an application (Data or VoIP). Depending on what you would include in the data center, this could be the router connecting the data center to the outside world/the company network, the firewalls protecting the data center, the distribution switches, the load balancers and of course the servers running the applications. In case of application performance degradation the question is who owns the escalation? Is it the Network, Server or Application “department”?
Solution: One integrated performance management system (Network, Server and Data-/VoIP-Application) that is capable of sorting out the escalation by pinpointing the problem domain.

Network Core – Different feeds and speeds: A key challenge in large core networks is not only the different speeds and feeds but also the fact that the size of these core networks has increased dramatically. This means there are more and faster links through which end users may access data on the data center server environments. Add asymmetric routing to that and the chaos is perfect when it comes to troubleshooting an issue in the core because the in and out traffic is not crossing the same link / Analyzer that has been deployed to sort out the problem.
Solution: Multiple analyzers with multiple cards supporting different speeds (100/1000/1000 Ethernet) and feeds (copper, fiber) possibly in combination with special tapping technology are fundamental to be able to troubleshoot a problem in the core. If you cannot see all the packets of a conversation, there is no way the problem can be seen, let alone solved.

WAN Management – Lack of visibility: The WAN becomes more and more of a “Black Hole”. The reason is MPLS. Carriers are migrating their networks to support higher traffic volumes as well as the need to reduce distances between branches and/or headquarters to support new technologies such as VoIP. MPLS allows the carrier to create a virtual network for each customer with what we call “any to any” connections. The positive effect is that headquarters traffic volumes drop dramatically. In addition, distances a packet (VoIP or data) has to travel between 2 sites could be reduced by 100% or more, which has a great impact on performance. The challenge is the visibility! In a traditional environment, where all traffic was routed through headquarters it was easy to understand what was going on on the network and to troubleshoot a performance problem. With “any to any” connections, the visibility from the central location is gone.
Solution: Depending on the depth of information needed, existing infrastructure information from the Routers (even from carrier edge routers) can be used to gain back the broad visibility. For the in depth analysis and troubleshooting dedicated HW can be deployed at key locations. All the information from the sources can then be managed in the integrated performance management system.

Remote/Branch Office – On site staff: Imagine a remote office or a branch office, that has no IT staff on site, a few hours away from the HQ. To gain visibility into the performance of the network and applications on that site is one thing. But to troubleshoot such a problem is a completely different story. Usually it means to send someone from the IT staff to the site with an analyzer to collect information which allows them to understand the root cause of a problem and to solve it. This is certainly not the most efficient way of dealing with a challenge like that. Apart from the cost and travel time for the IT person to go on site, the actual down- or browntime is a large factor.
Solution: Instead of have no visibility and/or having to send staff onsite for troubleshooting, there are options available today that allow you to gain visibility into remote sites for remote WAN/LAN monitoring and troubleshooting from a central location. The information collected from the sites will be centrally collected, stored and presented on the integrated platform allowing the isolation of problem domain and fast problem resolution without having to travel onsite.

External/Field Employees (VPN) – Performance & connectivity: Due to the nature of virtual offices with VPN connections using the public internet to access the corporate network there are and will always be 2 major challenges: A) The performance of the applications across the public network and B) the connectivity in general. A field with either of the problems will call into the helpdesk. First level support people will be challenged to help the customer if they don’t have any information about the connection of the user and its performance metrics.
Solution: The Helpdesk needs to be equipped with a system that allows them instantly to rule out an issue to the local infrastructure vs. the public domain. The presentation of the information needs to be in a way that even a first level support person can interpret the information. The next step here would be to provide the information directly to the customer online to help reduce the number of trouble tickets generated by public domain problems.

Corporate/Campus Office – Wired & wireless performance issues: Wireless has become a commodity nearly every corporate network today. For IT staff, the connections are really not an issue anymore. However, being connected doesn’t mean to have an acceptable performance level from the applications. High volume applications have a large impact on time sensitive applications. Meaning, if a high volume application “occupies” the air, a time sensitive application like VoIP may not get the packets in time to sustain a high quality level.
Solution: It is imperative to understand if the root cause of performance degradation is within the wired or the wired network. It does not help just to know that the network is the root cause anymore. We need to understand which part of the network is responsible for the issue.

Fluke Networks Visual Performance Manager equips your IT staff with a set of integrated tools to monitor and manage the core of your network, any branch or remote office as well as the connecting network in between to ensure the delivery of high quality application, VoIP and network services to your end users.

All teams throughout the IT organization will need unique views respective to their function. Regardless of user, any custom view can be created to meet each unique need. Cross-functional teams can easily see the problem domain, using the same data source and performance management system, helping to speed MTTR.

Some of the objectives a unified performance management system like Visual Performance Manager helps achieve the different levels within an organization:

Technical User

• Rapid identification, isolation and troubleshooting of performance problems

Director Level

• Introduce new business and application services faster, more efficiently and with less risk

Executive Level

• Integrated system that drives collaboration and empowers executives to optimize and sustain efficient business
  service delivery

Fluke Networks’s line of Network SuperVision™ solutions permits the testing, monitoring and optimization of enterprise and telecommunications networks, and the installation and certification of the fiber and copper forming the foundation for those networks. 

“IT staffs are very concerned with their ability to monitor and manage the myriad of products and services at their branch locations.  Often, they don’t have the tools they need to measure application and network performance before and after a new IT initiative.”