A History of SD-WAN

In the 1980s, in order to connect LANs that were in different locations, you used point-to-point (PPP) leased lines. These were typically DS0 (56 Kbps) connections, and later on the faster, more expensive T1/E1 or T3/E3 connections which could also be purchased as fractional  T1 or T3 lines at a much lower cost point.

Frame Relay service was introduced in the early 1990s. The same connections used with PPP could be used to connect to a “cloud” from a service provider. It was no longer necessary to purchase and manage individual links between each of the locations. Compared to PPP, Frame Relay reduced monthly WAN costs with far fewer physical connections to manage. It allowed the expensive last-mile link bandwidth to be shared across multiple remote connections, and used less expensive router hardware than the PPP. The OpEx and CapEx advantages of Frame Relay created an explosion of growth of the corporate WAN around the globe and within 5 years of its introduction, even the most conservative enterprises had migrated to Frame Relay.

In the 2000s, MPLS became the successor to Frame Relay and was designed as an IP-based solution for carriers to converge voice, video and data on the same network. Today MPLS, the most common deployment of enterprise WANs, is a connectionless protocol, whereas Frame Relay is connection-oriented. This difference gave MPLS an advantage with reduced latency in live voice calls and improved QoS.

In April 2013, the board at ONUG convened for its bi-annual meeting at UBS headquarters where use cases were shared requiring solutions that suppliers were not yet providing nor addressing. The ONUG Board, invited a handful of guests to provide their input and feedback including Jim Kyriannis, Program Director for Technology Architecture at New York University, who was the one to contribute to the “Branch Office Has Multiple Paths to Headquarters” use case.

It was at the following ONUG Conference, hosted by JPMorgan Chase, where the use case was again presented and its title was transformed into SD-WAN. The ONUG Community was asked to vote on nine use cases at that meeting and it was Jim’s SD-WAN use case that earned the vast majority of the community’s vote. The ONUG SD-WAN Working Group was launched and collaborated with 17 vendors on proof of concepts, including discussions about the cost, risks, benefits, and value.

As MPLS adoption grew, more organizations began to understand that MPLS had economic and technological advantages over Frame Relay causing a rapid migration to MPLS. Today, a similar shift is occurring as enterprises begin looking to replace MPLS with SD-WAN based networks. What has caused this newest networking technology shift? What are the prime differences between MPLS and SD-WAN which are motivating organizations today to look for another solution?

Most businesses rely on MPLS services for its dependability with SLAs that guarantee latency, packet delivery, and availability. In the case of an outage, the MPLS provider resolves the issue within a stated period of time or pays the requisite penalties. However, MPLS is not budget friendly in comparison to Internet services. According to Telegeography, in Q1, 2017, median 10 Mbps direct-Internet access (DIA) prices are potentially 1/3 less expensive than MPLS. The time it takes to order and install MPLS circuits is another factor in today’s fast-paced environment. Depending on location, provisioning can take anywhere from 3-6 months.

With the combination of growing bandwidth requirements and restricted network budgets, SD-WAN resolves the issues of cost and network scalability that MPLS presents without sacrificing the quality of service. SD-WAN offers the following advantages:

• handles a variety of connections and dynamically route traffic over the best available transport, regardless if that’s MPLS, cable, xDSL, or 4G/LTE.
• provides redundancy and more capacity using lower cost links with multiple connections at each location.
• measures the real-time transport quality (latency and packet loss) of each connection and applies Policy-based Routing (PbR) to route application-specific traffic over the most appropriate transport.

Bottom Line: the time of installation to delivery is far faster than MPLS. Some SD-WAN solutions offer zero-touch provisioning which allows the end-point to configure its connection to the WAN using the available mix of services at each location; a site can be brought online quickly without requiring a networking expert to be on-site for the install.

Technologies are born from the necessity to solve challenges that arise over time. The rise of SD-WAN was born from the changing enterprise environment and the need to adapt WAN infrastructure to meet these needs while staying within budget.

Projections from industry experts agree that the migration from MPLS to SD-WAN has begun and is continuing to grow rapidly. Andrew Lerner, Vice President of Research at Gartner, predicts “By the end of 2019, 30% of enterprises will have deployed SD-WAN technology in their branches, up from less than 1% today.” Another indicator is revenue from SD-WAN vendors is growing at 59% annually, Gartner estimates, and it’s expected to become a $1.3 billion market by 2020.

SD-WAN solution providers such as Cato Networks can help organizations make the transition and meet the challenges of today’s WAN environments. Subscribe to Cato Networks blog to find out the latest developments in SD-WAN technologies. To learn more, read more about the SD-WAN and SASE projections from Cato Networks.