Given that virtualization really is a paradigm shift for network architectures—how networks are developed and operated and the services we use are created, assured and billed—it is useful to remember that Software Defined Networking (SDN) has been around for a while. For example, did you know that WMFoundry was first introduced in 2000? Its close cousin for virtualization of CSP (News - Alert) infrastructure, Network Function Virtualization (NFV), has very deep roots going back to the soft switches of the early part of the 2000s as well.
In short, the network computing community has always sought to allow network administrators the ability to manage network services through abstraction of higher-level functionality. The reasons are as obvious now as they were when the concepts were first developed: speed, agility, control, performance and cost efficiency. And, at a high level, the engineers knew what they wanted to do to achieve these goals. It could be accomplished by decoupling the system that makes decisions about where traffic is sent (the control plane) from the underlying systems that forward traffic to the selected destination (the data plane). The rest is about leveraging APIs and virtualizing all aspects of operations.
It all sounded easy. However, like good wine it has taken some time, and we are not there just yet. Indeed, on the SDN side of things analysts are starting to question when and if the challenges relating to such things as the lack of real-time control, lack of equipment that can meet future needs and the need for further standards maturation have held things up. Fortunately, for CSPs there is better news on the NFV front.
One of the more interesting aspects of watching how both SDN and NFV are progressing, has been the question of whether overlay networks are up to the virtualization task, e.g., can they scale. In the category of what is old is new again, some recommended reading is the blog by Nuage Networks’ Dimitri Stilliadis where the answer is a definitive YES!
In what will be a two-part look at why that blog remains so relevant today, it is useful to examine what Stilliadis had to say about what overlay networks are and why they have been a vital part of the Internet Age. Indeed, as he notes we could not have gotten to where we are and cannot move forward without them. As he notes, “The core idea of an overlay network is that some form of encapsulation, or indirection, is used to decouple a network service from the underlying infrastructure. Per-service state is restricted at the edge of the network and the underlying physical infrastructure of the core network has no or little visibility of the actual services offered. This layering approach enables the core network to scale and evolve independently of the offered services?
Readers are encouraged to check out the entire blog. As Stilliadis explains Whether it is the PSTN or the Internet itself overlaying has been a core function to handling “E”verything relating to traffic flows and service delivery and assurance on wireless as well as fixed networks. He comments that: “Therefore, from a deployment standpoint, there is an extremely high probability that every time a user, whether consumer or enterprise, is accessing any network service, their packets are routed through one or most likely multiple overlays, that very often are not aware of each other.”
So what is old can in many respects be argued is new again. However, this is not about repurposing old technology. It is about harnessing the incredible pace of innovations in hardware and software to re-architect communications in a way that meets not just current but future requirements. It is why the questions about whether overlay networks can scale is so current and critical.
With the understanding of why overlay networks have worked, in the next posting we will examine what is “new” with network virtualization in the cloud environment. As a bit of a spoiler alert three questions will be addressed:
- Can overlay networking scale at data center capacities?
- Could a correlation between overlay and underlay networks improve the services offered in a data center environment?
- If the second question is true, are there open standards that can achieve this, or is there a need for some new protocols or technologies?
Stay tuned. The answers are certainly food for thought and remain highly important as questions about the future of SDN and NFV are a work in progress. The good news is that progress is being made.