|
|
|
|
|
Choosing Agility over Performance: Networking Strategies for Future Business Needs
Published Dec 9, 2014
|
The networks in use 10 years ago aren’t so different from those we use today, which is surprising given that the performance we expect from them now is exponentially greater than in the past. Whereas the IT manager’s greatest challenge was once connecting computers in different rooms, now thousands of employees in international offices expect to connect their desktops, laptops and devices to the network from their offices, from home, using Wi-Fi in public spaces, and using 3G and 4G networks while on the move.
In the past, the capacity and bandwidth of the datacentre put limits on the technology we were able to use for work, and issues with capacity could simply be solved by providing more bandwidth, with networks being the back-office conduit for accessing this resource. Now, the network is the beating heart of most businesses and absolutely vital to their smooth running. Without a network that is up to the task, coordination across local and international offices is impossible and the speed with which we are accustomed to doing business would be severely compromised. A flexible, efficient and intelligent network affords organisations greater agility, reduced time-to-market, increased reliability, and a serious reduction in the power and space used in the data centre.
Reliability is of particular concern as the demands of consumers on networks continue to grow, particularly where cloud computing has been adopted across many functions in the business. Data surges at peak times, which can put serious strain on the network and lead to choked provisioning, and the old remedy of adding bandwidth isn’t sustainable.
Simply throwing bandwidth at provisioning issues is no longer enough. In fact, this approach has in many ways slowed down the innovation and development of networking technology and strategy as organizations made do with a series of stop-gaps, rather than a long-term fix. That is now changing. No amount of bandwidth can prevent bottlenecks or latency issues, particularly not in the long-term with cloud computing now permeating much of our working and personal lives. As any IT manager knows, the more bandwidth you give a group of consumers, the more they devour. VoIP, videoconferencing and data-rich streaming, cloud software, and uploading and downloading of content put incredible pressure on today’s networks, and it’s a requirement that will only continue to grow. And grow it will – it’s predicted that 90% of existing data has been created in the last few years.
The drive for a holistic approach to provisioning, with the network at the core of IT strategy, is in part focused on new developments in networking standards and the innovative technologies they support, which will make networks flexible, responsive and intelligent.
The evidence to support intelligent networking is overwhelming. Google, for example, has been able to save millions of dollars every month since it doubled the efficiency of its networks. This was a huge investment, but one that was recouped in less than a year; and it will continue to pay dividends, as well as ensuring continuity of the network.
Network-Function Virtualisation (NFV) and Software-defined Networking (SDN) will be fundamental to most provisioning strategies in future and are hugely important developments in networking. IDC has predicted that SDN will be a $3.7 billion market by 2016[1].
SDN, which separates the control plane (the part that routes and directs traffic to and from the data centre) from the data plane (which carries the traffic), enables businesses to react to change by manually controlling the flow of resources, making network management simpler and faster. The real push for SDN has come along with the widespread adoption of cloud software and applications. Band-aid solutions in widespread use can no longer manage the pressures that cloud computing generates for networks, whereas SDN has the elasticity and scalability needed to cope.
By placing equal importance on the virtual and physical layers in the network, organizations will be able to build and manage networks that are able to contend – in real time and from any location -with the pressures of the future: reams of information coursing through networks at any given time compounded with data surges, the likes of which we have only seen hints of in the past.
The nature of the network an organization has in place often dictates the speed and manner in which they can adopt SDN. Simple, small-scale networks often seen in SMEs can more easily implement SDN across the network because of their size and a lack of complexity. They often tend to be heterogeneous, meaning they are not tied to a particular vendor. In these heterogeneous networks there is lots more flexibility and choice for IT managers to mix and match technologies that suit their current and future requirements. SMEs, through their use of SDN strategies, have created flexible, intelligent and efficient networks that act as the proof points for the technology’s deployment at enterprise level. For enterprise networks, particularly those with multiple data centres and offices spanning more than one country, it is less common for the network to be heterogeneous and therefore less likely that SDN can be quickly and easily implemented.
While it’s less simple to deploy SDN across an enterprise network where there are existing agreements with a homogeneous vendor, it’s still possible to use SDN strategically in targeted areas in the network. For IT departments in large or enterprise environments looking to implement networking strategies that include SDN and NFV, there are several roads businesses can take that don’t require a complete infrastructure overhaul. Often, IT departments start by deploying SDN in specialized places in the network, particularly when they have legacy equipment that isn’t heterogeneous and is tied to a specific provider. For enterprises, we imagine that within a year it will be easier to access the kind of ‘shrink-wrapped’ SDN solutions they will need to deploy it strategically across the business, however in the short term we’re seeing these small-scale, specialist deployments, often controlled within the hypervisor, which afford network reliability and efficiency where it’s really needed.
In years to come we expect SDN and NFV to become part of a wider, overall shift towards state-of-the-art, virtualized IP networks, which promise real savings on CAPEX and OPEX. What is referred to as ‘The New IP’ is an holistic approach to networking that involves a shift away from performance and towards agility and usability, in what is the biggest evolution in networking in over a decade. Ideal features that characterize a network following The New IP model include on-demand provisioning, centralization of intelligence and management, and software and user-centric operation.
Now that it’s clear that businesses cannot continue to add bandwidth as a solution to provisioning issues, The New IP is a framework for businesses to continue to grow at scale by using existing infrastructure as efficiently and intelligently as possible. Networks built on The Old IP simply don’t have the flexibility and scalability businesses need from their networks in the long term, and so these widespread changes in networking best practice are a natural step. In other words, what will set successful, cost-efficient networks apart from competitors’ in future will be their adherence to new networking standards created according to the current and future demands we place on our technology.
Posted by
VMD - [Virtual Marketing Department]
|
|
|
