The expression “hindsight is 20/20” couldn’t be truer for software-defined wide-area networking (SD-WAN). To summarize the past few years: Cloud computing and digital transformation drove companies to reevaluate traditional WAN technology, which no longer met their growing business needs. That’s when SD-WAN emerged as a promising new technology.
SD-WAN was designed to address the problem of traffic management from physical devices and to enable software-based provisioning from the cloud. Many initial SD-WAN deployments were fueled by the desire to replace expensive multi-protocol label switching (MPLS). Companies were hopeful that it could magically solve all their networking problems. But in practice, basic SD-WAN solutions fell well short on this promise.
Fast forward to the present, much of the hype surrounding SD-WAN has settled and early implementations are behind us. Now it’s time to look back on what we learned in 2019 and what to improve upon in 2020. So, let’s dive in.
1. It’s not about cost savings.
Most companies choose SD-WAN as a MPLS replacement because it can lower WAN costs. However, cost savings can vary with SD-WAN, so it shouldn’t be used as a primary driver for deploying the technology. Companies should instead focus on improving network agility—such as enabling faster site deployment and reducing configuration times—whatever their needs may be. The main driver for SD-WAN is to make networks more efficient, and if done successfully, the cost savings will follow.
2. WAN optimization is necessary.
Speaking of efficiency, WAN optimization improves the performance of application and data traffic. By applying techniques like protocol acceleration, deduplication, compression, and caching, WAN optimization can increase bandwidth, reduce latency, and mitigate packet loss. The initial notion was that SD-WAN could eliminate the requirement for WAN optimization altogether, but we now know that some applications need additional performance. The technologies complement—not replace—each other. And, they should be used to address different problems.
3. Security can’t be an afterthought.
SD-WAN has many benefits, one of which is using broadband internet to quickly send enterprise application traffic. But this approach also poses security risks since it exposes users and their local networks to the untrusted public internet. Security should be part of SD-WAN implementation from the beginning, as opposed to an afterthought. Companies can achieve the desired application performance and protection by positioning security close to branch locations using services like cloud-hosted security.
4. Visibility is critical to SD-WAN success.
Having visibility into application and data traffic takes the guesswork out of network management. The best place to start is at the pre-deployment phase, where companies can assess their existing capabilities and what’s missing before SD-WAN is implemented. Visibility, in the form of day-to-day monitoring and alerts, continues to play an important role after deployment. Companies that understand what’s going on in their network are better prepared to respond to performance issues and can use that knowledge to avoid future problems.
5. Wireless WAN isn’t ready for prime time. SD-WAN connects users to applications via any transport, including broadband and 4G/LTE (Long Term Evolution) wireless. That’s why mobile connectivity is increasingly being integrated into SD-WAN solutions. While companies are eager to use 4G as a potential transport alternative—especially in remote locations, the resulting costs of pay-per-use 4G services are high. Additionally, 4G can be problematic due to latency and bandwidth limitations. The best approach is to wait for service providers to deploy 5G with better pricing options. This will be the year when we see 5G rollouts and more focus on wireless SD-WAN.