The three main points to keep in mind for creating an agile network thanks to SD-WAN are:
1. SDWAN stays an important topic to address.
Beyond technology, the real challenge is to define the good “use case” and criteria for a successful deployment at a wide world.
2. Thinking about SDWAN needs to be integrated to a wider approach of the WAN and connectivity associated to it.
Regarding this topic, challenges are very clear (security, agility, cost reduction, cloud migration …), the issue is the way achieving it using WAN or not.
3. WAN definition is still on-going.
The boundaries are not clear between what we call “pure players“ (application prioritization, zero-touch provisioning, VPN) and editors for SDLAN and Firewall. This reinforces the need of identifying key functions for the WAN.
Let’s develop the first point regarding challenges to achieve:
a. Use Cases:
A rocket at two floors. Quick wins are identified (new site deployment, events, new applications with lot of data exchange, small website non protected …) but no global use case allowing transforming the WAN in SDWAND nowadays.
b. Best Practices:
Define the project planning and manage it on an agile mode. Planning aligned with end of MPLS contracts is still a good practice. During the RFI phases, retro planning must allow including the following points: quick win versus WAN from scratch, sourcing model (make or buy), organization impact, TCO consolidation, key functions prioritization, requirement review …
c. Solutions with variable dimensions:
Different architectures have been identified (uCPE versus SDWAN, MPLS versus full Internet). Choice is moderated by following factors: center size, geography, organisation (local/global).