The next step

Over the past couple of years, the corporate network has evolved significantly into a multi-service entity, with the convergence of voice, video and data services.

Now, as the economy improves, organisations are looking to new architectures to enhance business processes and provide a platform for cloud computing in the future.

With this in mind, their focus is on the data centre and the implementation of virtualisation technologies. Not only does virtualisation help reduce the equipment footprint, cut heating, cooling and other operating costs, it also represents the entrance into the cloud from a technological perspective.

Against this backdrop, virtualisation technology has become a key element of corporate strategic thinking. It is also an essential element of network consolidation driving ever tighter integration between the corporate network - including its switches and servers - and the storage network and its data repositories.

Companies are rushing headlong towards a time when switches, servers and storage devices will be consolidated within the form of the same virtualised network fabric.

While some industry watchers laud these developments, others are mindful of the increasing challenges facing network managers who will be called upon to manage this new virtualised topology, which will increasingly blur the distinction between the three elements.

One of the benefits of a virtualised environment is the ability to quickly move one machine from one physical platform to another physical platform. For example, network managers can shut one machine down, start up another while firing up a new R&D (research and development) system to understand and evaluate performance values and parameters. It's a very dynamic environment.

However, in order to effectively manage this environment, network managers will have to be able to identify the exact location of an application from a network point of view - as they did in the past.

Organisations are looking to new architectures to enhance business processes.

Andy Robb is chief technology officer at Duxbury Networking.

This will necessitate the ability to pinpoint the location - at any specific time - of a virtual machine on the network. Once its location is established and identity verified, network managers will want to apply the same security and other regulatory profiles to it that they did in the previous generation.

To realise their objectives, there will have to be an increase in the communication between the servers, the physical platforms and the virtual network infrastructures.

Of necessity, this process will have to be both dynamic and automatic. There'll be no opportunity for technicians to be constantly reconfiguring network switches to accommodate server movement!

But, will this be possible in a converged environment subject to the complexity of 'virtual machine sprawl' in which servers, switches and storage devices are constantly moving and repositioning between any number of physical platforms?

In the previous generation, switches and servers were part of the same physical platform and featured dedicated network cards installed allowing access, via a dedicated 'pipe', to defined storage devices often located on a separate storage area network.

Will communication between these devices - and their management - be negatively impacted when the corporate network and the storage network (together with its storage devices) become part of the same virtualised world?

One of the problems associated with older generation networks was poor operating efficiencies, in which jitter, latency and data loss were accepted facts of life.

These were (and are) particularly important considerations in the area of data storage, where 'data loss' is not easily tolerated at executive levels.

However, in the past, network managers were able to accommodate and obviate these disadvantages through a clear understanding of the physical network, including its line speeds at any particular point, and the specific, sometimes unique, needs of all the applications on the network.

In a virtualised network world, this ability will be hard won. It's predicted that the network industry will be looking for new means - new technologies - to evaluate and even forecast application performance across the network of the future.

Undoubtedly the industry will be called upon to drive currently competing global standards like iSCSI and Fibre Channel-over-Ethernet. Both are emerging IP-based networking standards geared to link data storage repositories and facilitate location-independent storage and retrieval over the virtualised network of the future.

The industry will also be called upon to fast-track the adoption of technologies such as Data Centre Bridging (DCB), a collection of standards-based extensions to classical Ethernet aimed at establishing the 'loss-less data centre' in a unified corporate/storage network fabric.

DCB pundits say it is reliable, offers predictable performance and can segregate and prioritise traffic. They say these benefits will encourage higher utilisation rates, simpler management and lower total cost of ownership - the same goals targeted by network managers of the past.