Convergence part one

From Network Computing Magazine Vol 14 No 01- Jan/Feb 2005

In the first part of a review of what IP Convergence is and can offer to an organisation and its users, Ray Smyth examines the landscape.

It is a fact that today the vast majority of company and organisational networks are not converged. This might for example mean that there is a network that carries data traffic, another which carries voice traffic, another that carries video traffic and – increasingly - another that carries security or surveillance data. These traffic types are not necessarily natural bedfellows, so why the apparent rush to converge them onto a single network infrastructure? The answer it seems is in the question – a single network; the migration of multiple legacy networks onto a single, integrated IP based network. The benefits of this complex equation cannot easily be generalised, but at the highest level, they will allow more efficient operational management, maintenance and general advancement of an organisations needs.

Convergence will bring much more than the Convergence of technology. It will help focus the service areas of an organisation, and a very obvious example will be between telephony and data. If such a Convergence offers scales of economy, this must manifest itself to some degree and eventually, in the Convergence of people and all of its consequences. We are even seeing the Convergence of departmental titles e.g. IT becomes ICT (Information and Communications Technology). Convergence may even mean less people, though it is a mistake to assume this.

Steve Kennedy of Thus, the network service provider says, “the converged network is becoming a utility”. An interesting view indeed when you consider that Thus is a spin-off from utility supplier, Scottish Power.

A typical organisation may purchase its various bandwidth requirements from a number of suppliers. For example, some inter office bandwidth for data and some for voice, some bandwidth to act as contingency for failure, and so on. All seems perfectly normal, however, if you aggregate the isolated bandwidth purchases and then overlay a utilisation analysis, it is highly likely that you will observe some huge chunks of under-utilised bandwidth across many network types such as Frame Relay, ATM etc.

Similarly, you might find periods of bandwidth restriction, which will directly impact service. In contrast, if all of these services were transmitted across a single network, it may be possible to reduce overall bandwidth requirements and improve utilisation of that which you purchased. Kennedy calls the disparate use of bandwidth a “Mesh Network”, explaining that “Thus is a converged Network supplier” and “offers a service that takes advantage of the ubiquitous presence of IP enabled devices such as routers”.

It is through the approach adopted by Thus, that you get to another vitally important fact and potential benefit of a Converged Network - traffic prioritisation. Standard IP offers no guarantees about its level of service or, for example, the order in which individual packets arrive at the destination or the timing of their arrival. This is not necessarily a problem and whether it is or not, will be determined by the traffic type. Data is highly resilient to the vagaries of IP service as it self heals and dynamically routes; packets can be assembled quite easily if received in the wrong order and at different times. The protocol will facilitate re-tries until it knows the packet has been received. In the case of, for example, voice packets, their order of arrival and timing is absolutely critical. As this shows, it is germane that a converged network is able to deal with different traffic types in different ways, reflecting their use and importance.

Multiprotocol Label Switching (MPLS) is one tool that can deliver this. For example where a layer 2 aspect such as latency is of critical importance to a traffic type such as voice (it is), MPLS will enable (through integration of layer 2 information and layer 3) such packets to receive priority treatment and thus attain an acceptable Quality of Service (QoS).

There seems to be an inevitable assumption that Convergence and VoIP (or IP telephony) is one and the same. It can be seen that this is not the case and that in a way, IP Convergence is a vehicle for the next stage of intelligent, and business focused, organisational networking. A vehicle that offers the next stage of operational value and organisational optimisation (a utility over which to deliver a complex array of services centred on IP) can now be more economically offered to the business. Add to this the merits of MPLS and you can really start to make use of the available bandwidth in a highly efficient and cost-effective way.

The justification for Converging your network may not be as cut and dried as you would hope. For certain, if you are installing a new network, installing cabling and hardware that will allow Convergence is assumed. However, the majority will already have a series of infrastructures already in place, all at different points in their capital write off cycle and usefulness to the business. This suggests that it might be reduced to a capital investment decision. Far from it, as some organisations will justify it from the dimension of the advantages it will offer to its business. One classic example of this would be a fully IP enabled call centre. Agents would become multi service agents and incoming enquires would be routed on availability and skills. So, an agent could find them dealing with a traditional incoming call, followed by a web chat, email or video session. Ian Shepherd of Telindus points out “that in general there are two types of organisation; those that see their network as a cost to the business and those that see it as a resource”.

Like everything pertaining to IT and Networking, a plan is essential. Unlike many other investment “opportunities”, Convergence is one where you can make investment and see no immediate benefits, let alone return. That said, I am convinced that 2005 will see all types of organisation making significant advancement on the road to high level IP Convergence. One driver will be the dearth of strategic investment since the Y2K fiasco – five years is at the start of a new investment cycle, though this alone will be insufficient to get board approval for such a strategy.

What of the supplier side – how are they responding to the Convergence opportunity? Unusually it seems to me that we are still seeing quite a lot of specialisation in the main component parts of the Convergence story. It is arguable that a few of the super size, usual suspects are covering many bases in their product offer, but I don’t see a huge rush to translate this into a rounded, service led offering. There are some interesting moves. For example, router supplier Juniper has announced a new range of products aimed at dealing with the high port density and reduced cost per port challenge, that must be resolved to fuel this sleeping giant of a market opportunity. Ken Cheng of Juniper says that “the choice is between perpetuating or converging systems”. Expect to see the price point move down.

It might seem that Converging onto one network might increase risk and exposure. Not so, as it is more accurately converging both physical and logical networks and a well designed network (as it always did) will provide resilience through its routing and MPLS priorities.

The convergence of legacy network services onto an intelligent multi-service IP based network is happening. Based on a sound understanding of your needs and some expert planning, a well delivered solution can offer many benefits and advantages. If your IT strategy for 2005 does not centre on a Converged approach, is your organisation disadvantaged?
 

Convergence part one

Uploaded by:
A. De Silva
Webmaster BTC Ltd.