We all know that the IT Infrastructure has a life of its own. In the vast majority of organisations the infrastructure evolves over time rather than being designed as a whole. This applie s to all of the underlying components: the servers, the storage and -an area very easy to overlook - the network or networks that tie everything together.
From a couple of decades ago, when Ethernet cabling first crept into businesses, the network infrastructure has slowly moved away from the concept of each system having its own cabling, routing, concentrators and modems and become far more generic.
With TCP/IP now a term recognised by one and all, the descendents of Ethernet reign supreme as the cabling infrastructure over which most systems traffic travels from servers to users. It is only in the area behind the server that links to storage that Ethernet has yet to sweep all before it. And this, too, may also be about to change as vendors look at various ways of using IP technology in their storage platforms.
As Freeform Dynamics has often commented, very few organisations have complete, accurate information on the IT systems they have deployed to support their business users. The lack of essential asset information is changing slowly as inventory discovery tools become simpler to deploy and use. And, much more importantly, as the value delivered by good asset and inventory management systems finally begins to be recognised.
Alas, when it comes to having detailed and accurate information regarding the networks deployed and the cabling systems in use, even rudimentary information is the exception rather than the rule. Do you know what you have in your infrastructure cabling throughout the business and the networking kit which keeps everyone connected?
Today, anyone surveying the market will notice that several vendors are beginning to talk about ’network convergence‘ with the idea that, ultimately, organisations will be able to use a single physical network that is essentially utilised to manage as many ’virtual networks‘ as may be required to service the varying needs of the business.
Clearly, in order to be able to reach such a state of nirvana, the organisation must have a complete picture of what is hanging off the network, who is using it and their service requirements. And such a picture must encompass a detailed understanding of the physical capabilities of the network components and the loads they are being asked to support.
This then brings forth the question of just how will such a converged network be managed? Come to that, just how are you managing your networks today? Do you have dedicated teams looking after the LAN systems, the WAN and the storage networks or is there a single team covering, as Douglas Adams would have put it, life, the universe and everything? Is each component being actively monitored and managed? Are all aspects of application service quality being monitored across the infrastructure? Our research suggests that this is very unlikely.
If all systems are to make use of a single network infrastructure then the application and service monitoring and management need to be a central facet of IT operations. This is, according to a good deal of research we have carried out, very much not the case today. Indeed, few organisations are in a position to dynamically manage and administer their networks in anything other than a few, very specific business critical instances.
Obviously, a few challenges need to be faced and overcome if organisations are to be able to exploit the benefits of IT service flexibility, enhanced reliability and the cost reductions that network convergence might deliver. Knowledge of exactly what they have in place today and how it is being used comes top of a reasonably long list.
No one is going to attempt to boil this particular ocean, so time will need to be invested to work out just where adopting such technologies might make most sense, with the storage networks perhaps being a good starting point for some.
One thing is becoming crystal clear – the management of ’the network’ is about to become much more visible and critical than it has been for some considerable time. All of which means that, for many organisations, it may well be time to undertake some strategic reviews of how these tasks are best undertaken.
Should they be combined internally into a single function encompassing the physical network management, application and service monitoring and management, perhaps coupled with some form of variable service charging? Are things best managed internally? Should some elements be passed onto managed service suppliers or moved to an outsourced network supplier? Or is the best approach to leave everything alone until a catastrophic service event forces a rethink?