If cloud adoption was the hot trend going into 2020, it’s now just a fact of life mid-2021. The way business is done has fundamentally changed in the interceding 18 months as people shifted to working from home en masse, and businesses were forced to become more digital at their core.
With this exponential increase in cloud computing and storage has come an increasing focus on the enterprise network. Businesses have moved quickly to try to ensure that they're able to continue to run smoothly and serve their customers, but far too often, they're considering just a portion of the network – that "last mile of connectivity" that data travels.
Often overlooked is what comes just before: the middle mile.
"Middle mile" as a term is actually a misnomer in itself as it gives a sense of equal weight with the last mile, but in reality, it is generally responsible for carrying traffic the majority of the distance across the internet. In the age of cloud, maximizing middle-mile performance is critical to providing excellent user experience and application performance. So why, then, is it so overlooked, and what can enterprise IT decision-makers do to ensure that they're looking at a complete view of their connectivity picture and achieving maximal network functionality and reliability?
Middle Mile 101
In short, the middle mile is the route of high-capacity IP backbone connections that data traverses at any given time as it moves from point to point around the world. The internet itself is made up of thousands of interconnected networks individually owned and operated by various Internet Service Providers (ISPs) – Tier-1 networks are generally the largest and most global, Tier-3 networks are the most local, and Tier 2s fall in between. The middle mile is the combination of all the multiple ISP networks that traffic crosses on any given journey from one end of the internet to another.
In an ideal world, these various tiers and networks play well together, and data flows smoothly. In reality, this isn't always the case.
Frequently, traffic must travel across a number of different networks across the spectrum of tiers. This creates challenges. Just as everyone is all too familiar with the traffic issues of an interconnected highway system, there are crunch points within individual networks and as multiple networks merge. How does each network route data? Where do the networks interconnect? What are the resulting congestion points both within and between the networks?
The more networks involved, the more questions arise, and the path becomes longer and more complicated. For example, it’s not uncommon for an enterprise connecting to a Tier 3 network to see its cloud traffic move to multiple Tier 2s, then a Tier 1, and at long last to its cloud provider. All these hops from network to network create a greater risk of latency which manifests in poor quality connectivity and unreliable, end-to-end performance.
MPLS to WAN and the Tier-1 network
The quality of the internet is especially relevant as many enterprises look to bring in cost-savings, increased agility, and cloud connectivity benefits by moving from traditional, hardware-based private MPLS networks to internet-based WANs. In order to ease this transition away from MPLS, businesses look to achieve better route management, capacity, and reliability – in other words, a way to exercise more control over their data's journey.
IT leaders in this position often do the simple math and come to the conclusion to select a single Tier-1 provider capable of providing a more direct route for their organization’s traffic. As Tier-1 networks are the largest and most global, sending data along this path from the start limits the number of hops it must take from network to network, dealing with transitions, different routing protocols, and a physically lengthy journey. In essence, the Tier-1 network gives the enterprise more direct control over the middle mile by eliminating unexpected challenges that lie outside of its ISP’s influence.
Further optimizing middle mile performance
While working with a Tier-1 ISP is critical to reducing uncertainty and ensuring network performance, there are additional approaches that can bolster the middle mile.
1) Find a provider with a strong Point of Presence (POP) roster. Point-of-presence facilities connect the ISP to the end user – or, in our terms, connects the middle mile to the last mile. A Tier-1 provider with well-distributed, global PoPs means more stability, reduced latency, and a smoother route to reach the end user.
2) Prioritize last-mile flexibility. While we look to limit the number of networks that data touches in the middle mile, a provider that's able to offer a broad range of local access partners is extremely valuable. These options allow IT and business decision-makers to have the necessary flexibility and agility to balance cost, performance, and risk needs, and tolerances on a per-site basis.
3) Embrace intelligence. In moving to an internet-based WAN, there’s great value in one that’s software-defined. SD-WAN technology helps intelligently and dynamically manage traffic and adjusts to network demands. Harkening back to the concept of the shortest, fastest route, SD-WAN ensures traffic is direct and efficient as it makes its way to the destination.
Middle mile and cloud performance
This focus on optimizing the middle mile is not connectivity for connectivity’s sake. It provides measurable performance improvement as businesses create an IT infrastructure that’s built on the cloud. A tier-1 network provider that exists at the center of that IT and cloud strategy offers simple but effective advantages of traffic flowing across a single backbone rather than hopping to various ISPs, reaching the proper PoPs depending on the physical path, and ultimately arriving at the end-user via the last-mile delivery service that’s right for the particular journey.
The middle mile touches every bit of traffic that will flow into or out of your organization. Doing what’s possible to maximize its performance results in low-latency connectivity and secure, private connections to the cloud service providers nearly every business now relies upon – or, in short, better enterprise network performance and user experience.
Michael Davies is Director of WAN Services Product Management at GTT.