In today’s fast-paced environment, CIOs are focused on saving money, enabling new revenue streams, increasing customer loyalty and creating competitive differentiation. At the same time, customers are asking for unique solutions tailored to support their specific communications and collaboration requirements.
Yet the data network, along with the applications and services running on it, are viewed more as a necessary evil than a tool to achieve these goals. The deployment of applications, stability of the network, and the time it takes to make changes can drag on for weeks or months or longer.
The network and its applications and services should lead the business, not lag behind it. In the end, enterprises must optimise worker productivity in order to become more competitive and more profitable while, at the same time, reducing cost and improving operational efficiency.
The data centre is at the heart of the enterprise and the transition to a next generation solution—one that supports needed capabilities—is top of mind within most organisations. In the past, enterprises were able to repurpose networking infrastructure equipment to fit the data centre since that solution was ‘good enough’ and specific requirements did not vary greatly from the campus. Now, ‘good enough’ is no longer good enough.
The concept of ‘fabric networking’ is a critical factor for quickly and easily deploying advanced business applications, whether or not they are virtualised. The enterprise data centre network must be flexible, less tree-centric and able to support sustained east-west flows between multiple servers and across storage, in addition to traditional north-south client/server transactions.
Factors driving the transformation of enterprise networks include the transition to composite application architectures, an adoption of business operations intelligence applications (big data), an increase in virtual machine mobility and a desire to converge storage traffic.
Each factor creates a unique challenge for the data centre network, ranging from security to latency/loss and increased traffic levels. What enterprises need is an agile, latency-optimised networking solution that delivers exceptionally high performance, reduces the total cost of ownership and is sensitive to the power and cooling requirements that plague the data centre today.
The dynamic data centre
The foundation of the new data centre is a virtualised networking architecture that enables enterprises to quickly deploy new applications and services while nearly eliminating human error. This new Ethernet fabric enables a private cloud—a virtualised data centre backbone which minimises complexity and delivers vastly more functionality and flexibility. Just like applications and servers before it, the network is in need of virtualisation.
An Ethernet fabric offers enterprises the ability to physically deploy servers and applications quickly and eliminates the multitude of configuration touch points required today. Network services now only need light touch or automated provisioning at the edge of the fabric to deploy new applications and services—there’s no need to redesign or make configuration changes to the core.
This simplified provisioning model greatly reduces time-to-service and the potential for human error by eliminating a significant amount of manual configuration. It brings carrier-grade reliability and performance at the scale and simplicity required for the enterprise. Seamless geographic redundancy can also be achieved since a fabric is self-aware, allowing it to automatically and instantaneously build a logical, optimised topology between servers, whether or not they are in the same data centre.
An increased level of security with inherent traffic separation and granular control over applications and services is another benefit of an Ethernet fabric. Segregation of traffic allows the network administrator greater latitude in securing access to applications and controlling the traffic within and between data centres. And by building it into the technology, traffic separation becomes the norm on networks.
Other critical characteristics for an Ethernet fabric to support includes Layer 2 virtualised networks, Layer 3 virtualised networks, routing internal to the cloud and dual-homing support for all servers, appliances and other devices attaching to the fabric. Lastly, a fabric must provide a migration path from the existing network architecture to the virtualised environment in the most non-disruptive manner possible.
Power and cooling in the data centre continues to be a concern as many installations are approaching maximum capacity. The move to virtualised applications can help significantly reduce the number of servers an organisation needs, helping ease the burden. The same concept can be applied to the network by moving to a two tier infrastructure and further optimising the network through virtualisation and the creation of an Ethernet fabric.
In the past, we saw disparate and competing networking technologies such as Token Ring, FDDI and ATM give way to Ethernet as the predominant network of choice – this provided opportunities to refresh infrastructures and create new networking paradigms.
Now, we are at yet another inflection point for networking as the move to an Ethernet fabric will radically change the way networks are built moving forward. This is a great opportunity for channel partners to become thought leaders and help customers transition into a new and unknown world where many of the ‘rules’ we have lived by to create networks are thrown out. By ushering in a new a way of thinking, channel partners can grow business opportunities and be on the leading edge of the next generation of networking technologies.