Over the past five years, we've moved from the software defined network (SDN) to SD-WAN (software defined Wide Area Network). In all cases, control of the network moves away from relatively fixed and inflexible hardware into more dynamic flexibility through software. The next step on the journey is moving beyond holistic local network boundaries with a single controlling entity, be it LAN or WAN, into the larger multiple party Internet. SD-Internet is coming and it's going to be interesting.
The SDN started out relatively small and humble in the data center, with control elements of the network getting pulled out of the routers and switches. Configuration on the fly for surge traffic or new processes became easy, with some keystrokes and a return, while also making it cheaper to manage and change things rather than locked into proprietary hardware operating systems.
SD-WAN is a natural evolution, extending beyond the data center and/or singular network into the relatively unique -- but larger -- enterprise space. Large businesses can optimize and more smoothly run networks containing multiple data centers, service providers, and distributed branches.
Moving software control onto the larger internet provides the opportunity for service providers, content providers, and businesses to define quality of service (QoS) in a meaningful way for all involved. With some exceptions, getting cloud services is a best effort based upon the network between a business, the content provider, and service provider(s) involved. Control of quality typically stops when traffic moves beyond the top edge of the service provider where traffic is handed off or exchange with other service providers. If someone drops the ball -- well, packets -- real time communications services are impacted and other cloud services become more difficult (slower) to use.
SD-Internet inserts a layer of intelligent control to the generic best-effort Internet, with the control layer provided as a service that constantly monitors and optimizes routing to more efficiently delivery packets between users and content sources, be it simple things like content and cloud-based Software as a Service (SaaS (News - Alert)) offerings or specific real-time applications such as conferencing and gaming.
At the level of intelligent control, real-time analytics enable the control layer service provider monitor and provide the fastest routing between two parties instead of relying on best-effort delivery between services and end-users. Businesses pay more for the intelligence and service, but gain faster quality and better statistics on what actually happens between themselves, their service providers, and end-users. Service quality is no longer an issue of "Well, after it leaves our Tier 1 provider, we don't have control" with SD-Internet.
For any enterprise moving or already depending on services from "the cloud," SD-Internet will become vital to ensure that services are not only reachable but easily usable, rather than being affected by third-party issues between content and cloud providers, the business, and customers.
Two particular technologies where SD-Internet will be useful are virtual reality (VR) and 5G. Real time VR requires plenty of bandwidth and computing power, so companies implementing cloud-based VR for engineering and gaming will need to ensure QoS between service provider and users. One of the basic features for next generation 5G wireless service is low latency. Wireless carriers and anyone wanting to take full advantage of delivering services over 5G will need a way to ensure that services are delivered as speedily as possible; old school best effort internet is going to be the weak link in adding sizzle and pop to multi-gigabit 5G broadband delivery.
Edited by Maurice Nagle