By Jim Poole
Continuing from an earlier blog post (Application-Aware Networks, or Network Aware Applications?) covering the CEWC 2011 debate on the relationship between Cloud and Ethernet, the discussion branched off into the idea that it’s not just network technology that needs to adapt to a world of cloud applications (or applications to networks), but the whole concept of how businesspeople think about IT.
Ciena’s Simon Parry explained his view: Thinking on SLAs needs to change: it’s no good keeping latencies and bandwidths being within agreed thresholds, but still having an unhappy customer because their cloud application performs poorly due to some peculiar problem with a specific application.
For many operators it might seem like their responsibility ends once they’ve tested latency and bandwidth and found them within agreed levels, but ultimately the application’s poor performance is often related to something within the network – the operator has to solve it, or they will lose customers to an operator who can.
This is a topic in itself, and had its own session with Olaf Herr from JDSU explaining how to test networks for application performance.
The panel was in accord that it’s not just the thinking of network operators that needs to change. The way business is done, and the way departments, lines of business and even individuals interact needs to change if cloud is going to realise its potential.
Colt’s Nicolas Fischbach said that in general, people’s attitude to cloud services was hardly visionary, and that both networks, cloud providers and end users were suffering from a piecemeal approach to adoption that fails to realise the main benefits.
A lot of blame was laid at the feet of procurement practice. The experience of the panel was that IT procurement was not generally carried out in a strategic manner.
Indeed, different departments tended to carry out their It programmes in isolation – often finding out from suppliers what other people in their own company were planning. Even at this point there is a reluctance to cooperate, generally due to difficulties in procurement procedure and funding activity across different parts of a business.
When cloud is adopted in tiny tranches, by different departments in a larger organisation, it’s not necessarily economically viable to create a fully application aware network through which to run these small adoptions, or to change thinking on established principles of SLAs to satisfy the demands of what is currently a smaller percentage of revenue than traditional deployments.
The networks and cloud providers can’t do it all on their own, if businesses are not looking at cloud deployments in a strategic way.
If cloud were adopted in a more holistic way, with agreements between business units or departments with similar goals, then bigger deployments could equal better networks and better applications at a lower overall cost per user. Everyone could benefit.
So perhaps the industry as a whole needs to work on changing people’s attitude to IT deployments. Once this mindset is reversed, the technology can be rolled out in larger, more effective deployments that will start to build up the infrastructure needed to fully realise the benefits of cloud applications.
That is not, however, to say that a successful high-performance cloud deployment is not currently possible, or even rare – but analyst figures show us there is still a lot of room for growth, and so perhaps there is merit in the panel’s consensus that a change in procurement and strategic mentality is just as important as any particular breakthrough in technology.