Turning Connectivity into a Competitive Advantage: the Age of Elastic Services Infrastructure
BY TOM GOWEN:
Cloud technologies have finally matured into an effective replacement of old Data Center and hosting methodologies. Some of those lessons learned need to be applied to the Networks that support them.
The Cloud has introduced many new concepts and innovative ways to design, deliver, and support global business applications, and among them, the one that soars about the rest, is the idea of a Hybrid Cloud model, or rather the idea that all applications are not created equal. What global companies like about this model is that Infrastructure should dynamically adapt to the needs of each application, doing away with convergence over a single (i.e. expensive) architecture, and doing away with manual configurations and static support of infrastructure. It also embraces the concept that in most cases the applications are not even hosted by the business anymore, rather by a Software as a Service or 3rd party provider.
While initial deployments of Hybrid Cloud solved the business need of choosing the right architecture for the right application, it was still missing the dynamic ability to automatically move workloads, or more importantly, the ability to automatically build services based on Infrastructure elements. Thus Elastic Computing was born. Elasticity in the cloud allows businesses to leverage virtual servers to enable services and apps to be moved around and scaled up or down on the fly without affecting the end user, saving time and money and enabling any business to be a dynamic global business.
The problem we face today is that while the Cloud is now smart enough to instantly change the infrastructure based on the needs of any global business, the underlying networks that connect these Clouds is not.
There has been much research and development on making networks more programmable, with software defined networking and the like, but little in the way of application to support the essence of the Internet of Things, which is really about a future that is highly interconnected, sharing data globally, and creating entirely new global business opportunities, applications and services.
To turn connectivity into a competitive advantage, you need to build an elastic network that’s flexible, adaptable, scalable and easy to manage, much like the Cloud has already done.
Similarly, elastic networks need to scale and adapt to match the ever-changing and unique demands on each companies infrastructure and business.
So what does Elastic Networking mean in practical application? There are many experts that have many different opinions, most of which are network hardware manufacturers whose opinions rely on their hardware to achieve. But from a conceptual view, the Elastic Network really needs to only have 3 things to be effective – Interconnectivity, Orchestration, and Technology-Automation. IoT could be easy to remember because it is meant to support IoT (Internet of Things)
Interconnectivity means the ability for the Network to talk to other networks inherently. With Business applications hosting in their own premise, 3rd party Data centers, and SaaS provider domains, the Elastic network needs to be able to have hooks into all other networks. Interconnection services like Amazon direct connect and MS Azure express route are only the tip of the iceberg in connecting to multiple 3rd party clouds.
Orchestration means that the network has to be controllable from a portal or integrated orchestration engine. This not only enables devices and services to be created on demand, but it truly integrates the network elements with the cloud compute elements. It allows business true abstraction of infrastructure. No longer does an application guy have to ask an IT guy for a server with a load balancer and a firewall and an internet connection, he simply presses the “create apache server” button and the orchestration engine dynamically builds the underlying infrastructure.
Technology-Automation means that the network needs to also support on-demand creation of virtual network service functions or traditionally called Network Functions Virtualization (NFV). This gives the ability to not only create virtual network services, but in the long term, allow the network to intelligently deploy such services without human interaction. (i.e. if a site gets hit with a DDoS attack, the network can automatically deploy a scrubber service as needed and then disable after the attack, giving the customer a dynamic service that he only pays for while in use.)
To achieve this dream of Elastic Networking, there will be much trial and error, much like in the early days of Cloud adoption, but for a start, to partner with service providers like Global Cloud Xchange, who has already begun to Interconnect our large capacity backbone to multiple 3rd party clouds, has created a scalable orchestration engine that will soon interconnect network services, and NFV services coming to our Cloud X IaaS platform soon, you will be one step closer to a smarter network that gives you an advantage over your global competitors.