S
ervice providers are facing real challenges with the explosion of data and devices bearing heavy on existing networks. With traffic volume increasing and becoming more diverse, many are now scrambling to keep up: fragmented networks with poor visibility are making matters worse.

For many, the answer so far has been to overbuild the network, whilst keeping it running at half capacity to accommodate any spikes in traffic. The result, of course, is a lot of excess with capacity sitting idle.

Service providers need the right tool kit to handle this increased demand – the key component of which is global visibility to enable reliable forecasting and advanced or real-time optimisation. The network needs to be run more efficiently, with overprovisioned capacity reduced and control exerted via customisable, multiple-layer management.

The Problem

The problem is compounded by a certain reluctance to tackle these issues proactively. This is understandable: the core network is business critical. If it’s functioning, why touch it? The risk can seem too high to justify the reward. Many of these cumbersome, legacy networks are left ‘as is’, because they appear to be ‘good enough’, for now.

In that environment, it can be incredibly easy to overlook the provisioning challenge at hand. Automation, and the automated provisioning of workloads, tackles this problem as it removes the need for human intervention in device configuration. A template approach, executed through tested and error free programs provides a scalable and repeatable solution to the challenge of manual intervention. As such, optimisation will assist in the provision of new services, whilst preventing any impact on existing SLAs.

The real challenge however remains network traffic. From both a volume and diversity perspective, it is on an unstoppable upwards trajectory. Now is the time to lay the foundations of a scalable, flexible network that can serve the next generation of modern enterprises.

The Solution

The future is here. The enterprise demands more from the network than ever and with bandwidth becoming increasingly commoditised, service providers have to find other ways to add value. Each of these challenges point to one solution, an autonomous network that self-optimises traffic flows in real-time, with global visibility, to enable human decision making with certainty. Efficiency must become king, through WAN SDN controllers that allow service providers to customise services with agility, and deploy them consistently.

Conclusion

The solution lies in moving from the ‘old way’ to the ‘SDN way’. That means transitioning fully from human workflows for network design and planning, to autonomous network optimisation. By breaking down the ‘old way’ into stages, it’s clear there are inefficiencies that need to be overcome.

In an offline, human executed workflow, resource is required to pull analytics, topology and the network configuration. Then further human resource is needed to simulate the network environment, assess trends, analyse them, decide how to optimise the network from the learnings and deploy new provisioning templates. This process must then be repeated for every cycle change or new service deployment to keep the network optimised.

The ‘SDN way’, on the other hand, frees up resource and associated costs by making the network more intelligent and autonomous. With a decision-making network able to self-optimise, service providers can enjoy:

• Online design, planning and monitoring
• Real-time analytics and topology discovery
• Continuous network self-optimisation
• automated provisioning of new services and other changes

This requires north-bound APIs for path provisioning, path computation and topology discovery, to allow custom applications and integration with self-service portals

By deploying a WAN SDN controller, capacity upgrades can be targeted, service driven and dynamically provisioned, to meet customer demand in real time. We partner with Juniper Networks to provide automated, scalable and secure network solutions that offer agility, performance and value.