Everyone deploying SD-WAN talks about making networks fast, reliable, simple to deploy and manage while maintaining high quality and performance levels. But how do you choose the right platform for your needs? Understanding some of the inner workings — like the difference between session-based vs. packet-based load balancing — of SD-WAN can help you during the selection process.
One of the key drivers for whether an SD-WAN solution delivers the expected (and desired!) performance is the difference between session-based vs. packet-based SD-WAN — how traffic is distributed over circuits in the tunnel. The way these different platforms transmit data can have a significant impact on the quality of the end-user experience, the speed and quality of applications, and the reliability of the network.
Let’s take a look at the major differences between session-based and packet-based SD-WAN and examine the impact each has on the user and network experience.
Session-Based Load Balancing
SD-WAN platforms that rely on session-based load balancing transmit each user application session over a specific circuit in the SD-WAN tunnel. Session-based platforms assume that bandwidth is plentiful, that circuits in the tunnel are always available and that network conditions are predictable and consistent. Based on these assumptions, a session-based SD-WAN isn’t designed to manage the tunnel for efficiency.
What Does That Mean?
If a circuit in a session-based SD-WAN fails, flaps, or experiences significant packet loss, the session is restarted on a different link in the tunnel. This typically creates a temporary outage for the user.
Packet-Based Link Load Balancing
Packet-based link load balancing treats all packets originating in the LAN as individual and unique. The SD-WAN software in the CPE at the end of the LAN decides which circuit in the tunnel to use for each individual packet. SD-WAN platforms that use packet-based load balancing are designed with the view that networks (and network conditions) are variable. This is critical when the SD-WAN includes broadband or wireless circuits in shared or over-provisioned networks.
What Does That Mean?
A packet-based solution enables the SD-WAN application to respond to degrading circuits by steering packets quickly away — and onto circuits that are performing better. This can be done without a change in IP address, so disruptions to the user are minimal.
Packet-based link load balancing works even better when complemented by QoS engines that are dynamic and allocate network resources to priority data flows only when priority packets are detected, and maximizing bandwidth when priority packets are not present. This approach can improve network efficiency and performance without requiring increased bandwidth.
Session-Based vs. Packet-Based: Which is Right for You?
The decision of session-based vs. packet-based load balancing really comes down to philosophy and use case. From the perspective of efficiency and elegance, a packet-based approach wins. It respects network bandwidth and can accommodate a wider variety of use cases. In contrast, session-based approaches are easier to develop and outperform other wifi or security vendors’ technologies where SD-WAN is treated as a value-add to the main product line.
When choosing between session-based vs. packet-based load balancing, consider your ultimate goal. If the main goal of deploying SD-WAN is to keep all solutions within a single vendor’s product line, a session-based solution will work. On the other hand, if your goal is to deliver a best-of-breed solution, integrating a standalone packet-based solution with existing or new security or wifi management options is probably the best option for you.
Are you struggling to choose between session-based vs. packet-based load balancing? Want to know more about how to get the best SD-WAN solution for your organization? Contact the experts at Multapplied with your questions or schedule a demo today.