Opinion: Wayne Sanderson, Head of Networking Business Unit at Dimension Data, looks at the importance of an effective data centre network.
As a business grows, it is essential that its data centre network evolves with it to achieve increasingly demanding performance goals. Failure to do so can be potentially devastating and result in the business flatlining.
The faster, more agile and more responsive data centres are these days, the more aware, intelligent and competitive the business becomes, however, achieving this requires a holistic approach that involves the data centre network evolving at the same pace as everything else.
With growing pressures to reduce carbon footprint and organisations always looking to reduce costs, it’s no surprise we have seen a trend towards more consolidated data centres. Data centres are increasingly being expected to deliver more capacity at greater speed, while consuming less power. The extent to which they achieve this is often down to the sophistication of the server in place.
New servers, in terms of their microprocessors and memory, are indeed far more powerful and efficient than ever before. However, it’s imperative not to forget about the underlying network that brings them together.
It has often been the case that an organisation choosing to deploy much more powerful servers within its data centre will then find itself needing to immediately upgrade its network capacity in order to keep up with the increased speeds. A server refresh is often the impetus for organisations to start thinking about evolving their data centre networks, but there’s more to it than simply considering hardware and capacity requirements.
Fragmented applications, enterprise mobility and cloud computing
Along with a change in sophistication for servers, the applications that run within data centres have also changed. The days when a single sever would deliver an application are long in the past – nowadays, it’s not uncommon to see three-tier applications. A single server handles the interface to the user but it is connected to a database server and web server in the back-end. Different parts of the applications do different things.
This ‘meshwork’ of functions behind the scenes results in large amounts of what is known as ‘east-west traffic’ where information is passed back and forth between various physical and virtual machines while users experience a unified view. Architectures that have a higher level of flexibility are, therefore, required to meet these functional and structural evolutions.
The shift, in today’s culture, towards enterprise mobility and cloud computing is further adding to the pressure that data centre infrastructures are under. Users’ demands have reached the point where they now expect to be able to connect to applications from anywhere in the world, instantaneously, and have the same experience regardless of the time or their location.
The flexibility, scalability and pay-per use benefits offered by cloud computing are ideal for any business if it can be seamlessly integrated into legacy infrastructures. However, because we make so many demands of data centres today, without an end-to-end review and transformation of architectures, organisations will run the risk of flatlining under the pressure.
The network is the platform for an organisation’s entire ICT infrastructure
This holds true within the data centre. Modern data centres and modern applications can’t function optimally without modern data centre networks. Without this foundation to build on, the whole structure becomes weak and can have a detrimental effect on the organisation as a whole. When addressing this, it is important to consider each of the following foundation elements of the network.
It is crucial to have the right hardware in place, especially now that specialist data centre network equipment is emerging. A general-purpose network switch simply won’t work. What is needed is data centre switches that offer a set of special features targeted at the specific environment. Front-back airflow is one example, designed so that the device fits properly into a rack. Another is the ability to run storage traffic, which enables converged networks inside the data centre.
In addition to specialist hardware, an important evolution in networks overall is the emergence of software-defined architectures, which is particularly useful within data centres. Software-defined data centre networks are highly automated, eliminate human intervention, and add to the data centre’s overall programmability, agility, and flexibility.
Specialist network functions are now also run in virtualised software environments which allows for a great degree of control to network management.
When collectively applied, it is these foundation elements that come together to create the structure of the modern-day data centre network. In addition to these building blocks, technical aspects such as application acceleration and network services may also come into play to ensure that users accessing data within the data centre have the speed, reliability and quality they desire.
In order to achieve the architecture that will serve your business in the best way possible you need to have a clear view of what it is you want to achieve with your infrastructure, and have your technical and business objectives in mind. Migrating towards that end-state can start only once you’ve created a step-by-step roadmap.
Consider all the various benefits and pitfalls of the individual elements that go into building your ideal data centre network. It isn’t necessarily a case of rebuilding your entire infrastructure all at once. Sometimes, in order to gain experience and confidence, a specific area can be focused on. Remember – it is all about executing an appropriately paced transformation over time, which also implies evolving your data centre network at the same time.
If your data centre network transformation is ignored, or executed haphazardly, it can lead to stunted business growth through a failure to meet demands.