By Patrick MeLampy
The SD-WAN market is exploding – with some estimates projecting it to reach as much as $9 billion in the next four years. While the proposed benefits seem impressive – improved agility and cost, enhanced security, and automatic provisioning – SD-WAN technology will prove to be little more than a short-term solution. This is because mainstream approaches to SD-WAN follow the standard practice of solving a network problem by putting a solution on top of what’s already there, essentially overlaying a network on top of a network and creating more complexity. Ultimately, this fragile patchwork of technologies will not be able to withstand the increasing pressures placed on them by cloud, mobility, security, and application proliferation.
Considering the unique challenges cable operators are facing, this idea is hardly revolutionary. As they try to sell to larger businesses with national footprints, cable operators are finding that they need to partner with competitors to reach branch offices outside their service area. While this may sound simple, it is not.
Currently, 30 to 40 companies are making SDx products. Virtually all of these companies use overlay network techniques and some form of centralized control. These techniques work well when a single operator can manage the service. However, when multiple operators are involved in providing a single service, these approaches break down. This is because overlay techniques require coordinated provisioning to work. If they are also performing cryptography as part of the solution, then automated key management is required. Performing this level of coordination across operators is quite difficult without standards and protocols.
Currently, SD-WAN vendors offer proprietary approaches that do not interwork with each other, meaning that one SD-WAN company’s solution cannot connect, communicate, or exchange data with that of another. The lowest common denominator to which each of these solutions can talk is the IP network that sits below these solutions and provides the real network.
For cable operators, this real network is typically a flat network that shunts traffic to the internet in a distributed fashion. This keeps their backbone networks small, and is used primarily for delivering services. When trying to offer business services for companies, cable companies can either build out a core network much like a traditional data networking company, or they could use smart, multi-path routers that support secure vector routing (SVR).
SVR is a technique to connect two branches of a private company across a public network that allows a “flat” network to provide VPN-like services with end-to-end encryption and improved connectivity. This technique works across multiple operators, and does not require coordinated provisioning – just IP routing protocols to work, drastically simplifying the structure of the network. The current routing protocols are sufficient if they are augmented situationally through an open standards-based exchange protocol with service and topology information.
There can be no question that the real network supporting this web of overlay solutions needs to be fixed. Currently, it is a patchwork comprised of complex solutions that do not address the larger problems posed to the internet. To overcome this, cable companies need to look beyond the SD-WAN hype and instead towards next-generation IP routers to meet the current and future demands facing their customers.
By fixing routing and using an SVR approach instead of relying on overlays, the real network could perform multi-path routing natively, support networking end-to-end through multiple network borders, and remove these disparate technologies that prevent cable operators from offering the seamless connectivity and flexible services their distributed enterprise customers need.
Patrick MeLampy is the Co-Founder and Chief Operating Officer at 128 Technology
The original article can be found here.