The “last mile” creates a unique set of challenges that has plagued industries ranging from transport & logistics to urban planning to old analog POTS lines. Like these other domains, enterprise WAN has struggled with last mile optimization as well.
In a nutshell, last mile problems are the inefficiencies that occur in the final leg of delivery of a service. For WANs, this is the last stretch of connectivity from a service provider to an end user (e.g. from an ISP to a remote office). With so many variables to account for, mitigating downtime and optimizing performance in the last mile has proven to be a difficult problem to solve.
Fortunately, with the advent of SDWaaS (SD-WAN as a Service), enterprises now have a viable global solution to last mile problems. By providing a robust and converged network infrastructure with advanced monitoring and HA (high-availability) functionality, SDWaaS enables enterprises overcome last mile problems on the WAN.
In this piece, we’ll dive into some of the specific challenges the last mile poses to the enterprise WAN and explain how SDWaaS is able to address them.
What kind of problems occur in the last mile?
So, what makes the last mile unique? The answer becomes a bit clearer when you understand the “middle mile” and think about how data makes its way from point A to point B.
The middle mile is the underlying network infrastructure that accounts for most of what the ground data covers. Service providers can have a fair amount of control over the performance of the connectivity here, and in many cases, it can be inherently robust. If one “hop” along the way goes down, there is often another path for the traffic to take.
However, in the last mile, you are subjected to the quality of the connection between a given endpoint and the ISP (Internet Service Provider). If that link has an issue, there isn’t really an alternative path for the data to take. If a squirrel chews through a line, a router goes down at an ISP, or provider’s network is simply overworked, getting reliable network performance for a given site becomes almost impossible. This becomes an even bigger challenge at a global scale when branch locations may operate in areas where connectivity isn’t great to begin with.
The end result is enterprises are stuck dealing with a number of connectivity issues for specific remote sites. Packet-loss, brownouts, and jitter become the norm and enterprises do their best to mitigate them.
How SDWaaS solves last mile challenges on the enterprise WAN
So far, it seems like enterprises just have to play the cards they are dealt when it comes to the last mile, right? Unless they’re going to build out the ISP infrastructure themselves, the best they can do is build redundancy into their networks and try and architect their way around the seemingly unavoidable last mile challenges. However, it doesn’t have to be that way, and this is where SDWaaS comes in.
One of the major advantages of SDWaaS in the last mile is that the provider has Points of Presence (PoPs) across the globe. Thanks to the strategic placement of these PoPs, SDWaaS providers can provide enhanced connectivity in the last mile. These PoPs also make up the SLA-backed global backbone that has helped earn SDWaaS a reputation for reliability. Below, we’ll dive into some of the specifics of the benefits of SDWaaS in the last mile.
Reducing Packet Loss
Lost packets can be a nightmare problem, particularly when dealing with applications such as VoIP or UCaaS. SDWaaS helps address packet loss in two ways:
- Intelligent Forward Error Correction (FEC)– Intelligent FEC leverages an algorithm to scale up or down the amount of correction packets depending on the number of packets lost and the capacity of a given link. Further, it can be configured based on the requirements for a particular protocol, port, or app.
- Packet Duplication– FEC isn’t always enough on highly unstable connections. Packet duplication sends packets across both links in an active-active connection for some (or, if desired all) apps. This feature is particularly useful for apps that are highly sensitive to packet loss (e.g. aforementioned VoIP or UCaaS) or where the quality of the last mile connection is extremely poor.
Improving last mile resilience
Redundancy is an important part of last mile resilience. SDWaaS enables the connection of multiple links and the ability to route traffic on two links in parallel in an HA (high-availability) active-active configuration. Constant monitoring of the links allows the SDWaaS network to leverage the connection with the least packet loss.
While many WAN solutions, SDWaaS included, can help address blackouts, detecting and responding to degraded performance, a.k.a. a brownout, can be more challenging. SDWaaS offers brownout mitigation that detects poor WAN performance (e.g. high packet loss) and switches to a different link. The dreaded “flapping” that can occur with unchecked failover is mitigated by configurable interval settings.
Reducing Latency & improving throughput
SDWaaS also delivers traffic shaping and performance enhancing features such as Quality of Service (QoS), bandwidth throttling, and Policy-based Routing (PbR). These features are an important part of prioritizing the performance of mission-critical apps and services.
Further, SDWaaS PoPs can act as a TCP proxy server. This leads to TCP clients and servers potentially using larger TCP windows and helps enable TCP congestion control. Additionally, dynamic PoP selection helps to ensure that the PoP with the lowest latency and packet loss is used to route traffic. Real-time updates help ensure that the network automatically selects the best PoP and delivers the best performance possible.
SDWaaS has cracked the last mile problem
The last mile has always been a difficult problem for WAN engineers to crack. With legacy WAN solutions like MPLS, engineers were at a disadvantage and limited in their ability to mitigate last mile problems. SDWaaS offers enterprises a new approach to WAN connectivity that leads to significant improvements in the last mile. By coupling a robust global backbone with failover, monitoring, and performance-enhancing features, SDWaaS serves as a reliable way forward for the enterprise looking to reduce the challenges of the last mile.
