Jan 17, 2013
So here we are — 2013 already. No matter what New Year’s resolutions you have made, can I advise a few more to take on so that your organization is ready to better face what promises to be a tough year ahead?
Firstly, review your current approach to data center networking. Over time, it is likely that the data center network topology has morphed into a bit of a monster — hierarchical systems of switches leading to complex routing of packets of data with all the associated latency issues that this can lead to. Moving to a fabric-based approach can flatten the network, reducing not only the latencies involved with East-West traffic of different enterprise applications trying to share data between them, but also helping in the North-South traffic between the data center and the various types of devices being used to access the applications. Bringing in top- and end-of-rack switches where the network can be virtualized gives far greater control over what is happening at a transport level.
However, this may not be enough, and so secondly, you should look at Software-Defined Networks (SDN). SDN provides a means of abstracting many of the network functions from the proprietary environments of network operating systems, providing a more standardized capability to deal with areas such as security, availability, load balancing, and so on. As these functions are moved away from dependencies on any specific vendor, SDN also enables a far more open network hardware approach, and also allows for advances in networking functionality to be more rapidly adopted without the need for forklift upgrades at the hardware level.
Thirdly, prepare for the data center to move further away from being the center of your organization’s overall IT platform. With mobility and the rise of BYOD (Bring Your Own Device) meaning that employees are using a mix of devices in a range of ways, with the network connectivity outside the direct control of your organization, it is increasingly important to ensure that access to applications is maintained (wherever your employees are). This may mean that a more strategic approach is required for the use of external networks – multiple, redundant connections from the data center to the internet and multiple wireless and broadband plans for employees so that they can be pretty much assured of connectivity at all times.
This will become even more important as organizations continue to embrace cloud computing. An increasing amount of functionality will be provided from outside your own data center, and ensuring that these disparate and decentralized bits of functionality are all available and performing well should be an imperative. So fourthly, you should ensure that the means to measure end-to-end performance are in place, to report in a predictive manner the trends and issues users are having, and to be able to deal with any issues arising in as automated a way as possible so that users remain unaware of any major problems.
Such application performance monitoring (APM) tools are available from many vendors, as are tools to manage performance across multiple different external networks through load balancing, caching, packet shaping, compression, and so on.
Globally, 2013 is likely to remain a tough year. Consumers are struggling with their finances and this will feed all the way through to business-to-consumer (B2C) and business-to-business (B2B) organizations. The past couple of years have been more around the survival of the fittest — 2013 has got to be more than this. It will be a race beyond survival to the defining of the winners in vertical and geographic markets. Much of this will be predicated on how effective an organization’s networks can deal with the pressures placed on them. Prepare now — and plan to win.
Image credit: Phil Roeder (flickr)