While the last 10 years have seen network functions consolidated into a single physical appliance, the move from the physical to the virtual world is now producing the opposite effect. And all signs suggest that will accelerate further in 2016. This decomposition of the network will see more network functions migrate back to their constituent parts and this trend to network function virtualisation will increasingly change the way networks are designed and managed.
Virtualisation removes the boundaries of where network capabilities need to run, so that functions can be deployed where its makes more sense. For example, since virtualised network functions can be placed closer to the end user, compute power can be distributed around the edge of the network, reducing the need for packets to be routed all the way to the data centre, which is critical for low latency networks.
Another example can be found in load balancing. The capabilities of traditional server load balancers have been extended to include WAF (Web Application Firewalls), IPS (Intrusion Prevention Systems), firewalls, SSL accelerators, compression and caching, and have become known more commonly as Application Delivery Controllers (ADCs). But what makes sense in the physical world no longer makes sense in the virtual world. In the ADC use-case, we might now place the firewall or IPS at the edge, while keeping load balancing capabilities close to the application within the data centre. This will prevent so-called ‘bad’ traffic from having to traverse the entire network and ultimately off-load the data centre.
As well as the trend towards greater network decomposition, here are some other trends I expect to see emerging further in 2016:
Business adoption of ‘Internet of Things’
While the connected fridge and toaster are still well away from the mainstream, innovative businesses are now embracing the IoT to improve information gathering and analysis and as the foundation for automation and control. The scope for IoT deployment will be increasingly applied to large and complex entities such as manufacturing plants. This evolution will drive not only an increase in data volumes and sources, but also require additional layers of security and application delivery. These additional layers will be delivered in a virtualised environment as dynamic microservices that can adapt to changing usage patterns that often characterise IoT services.
The Software Defined Data Centre (SDDC)
The need for IT services to quickly adapt to support business dynamics will lead to the adoption of orchestration and automation toolsets to simplify IT operations. Taking a software defined approach enables IT to react quicker, make better use of resources and have a unified view of application delivery from the network through to application components. Key to a software defined approach is the use of the emerging technologies of software defined networking (SDN) and network function virtualisation (NFV) in combination with well-established server and storage virtualisation technologies. Having an environment that is virtualised from the application at the top to the network at the bottom will be the key enabler of the software defined data centre in 2016.
Emergence of Adaptive Security models
With the emergence of more complex and dynamic application delivery infrastructures, traditional security tools designed to identify known threats are insufficient and we will see a more adaptive approach to security. This model will see security policies and behaviours ‘baked’ into the application delivery stack, which will drive a fundamental change to how applications are developed and deployed. The capability to deploy and orchestrate security functions as microservices enables the application of security policies at multiple points in the estate rather than the traditional approach of deploying at the network edge.
Use of APIs to drive rapid development and deployment
The DevOPS model for software development and delivery will drive the use of third-party tools and microservices to reduce the time to delivery. The key to the use of such microservices is the widespread use of APIs that not only provide the application functionality but also enable management and orchestration. In the DevOps model, being able to provision, monitor and control via APIs is essential as it is automation that enables continuous delivery and ensures consistency. With API driven microservices providing network and security functions, DevOps environments will be able to take a Lego type approach to creating and deploying services in 2016.
There is no doubt that our industry is guilty of too much hype and over enthusiastic predictions about the future. But there is a sense of real change that is here to stay driven by some exciting new technologies and innovative approaches to network design and optimisation. This gives me the confidence that these trends will gain pace in 2016 and become mainstream over the next few years.
Jason Dover, Director of Product Line Management at KEMP Technologies