How Section Achieves Greater Scalability with Edge Infrastructure Decisions

November 12, 2018

At Section, we’re primarily focused on providing value for developers on the programming side of edge compute, but the scalability of our platform is heavily dependent on the edge infrastructure that we utilize. In order to meet the demand for increasingly vast volumes of requests that pass through our platform, we’re constantly looking for ways to optimize our infrastructure to continue to support highly performant systems.

One of the biggest advantages of our edge computing software is the flexibility and control that it offers for developers to run any workload, anywhere. In other words, everything that we do from a systems architecture standpoint revolves around the freedom of choice that we extend to clients to be able to leverage edge computing within their own server and network infrastructures, from cloud to on-premise, and in-between.

As Section continues to attract, onboard, and support larger clients, we’ve been enjoying greater flexibility and control over the physical infrastructure that supports our platform. Each infrastructure decision for us and our clients, from providers such as AWS, Azure, Google Cloud and Digital Ocean, offers unique advantages and serves a specific purpose, and it’s the combination of those services that ultimately determines performance, efficiency, and scalability.

We’ve recently been hard at work implementing a large-scale platform upgrade to support our growth, requiring us to make critical infrastructure decisions to meet the growing demands. One such decision that has been particularly noteworthy is our engagement with Packet, which has allowed us to scale and expand flexibility for our clients.

Packet is a bare metal cloud, offering a fresh approach to cloud computing by providing on demand access to dedicated (e.g. bare metal) infrastructure all over the world, including at the network edge very close to end users. Packet’s proprietary technology automates physical servers and networks without recourse to multi-tenancy or virtualization, which means customers ‘own’ the hardware they deploy on the cloud, allowing them to use all the resources of that hardware.

Infrastructure to Support Edge Workloads

Section leverages Packet’s bare metal as a high-performing infrastructure foundation, layering our programming interface on top which gives developers access to the edge.

When developers understand that instead of scaling up their centralized infrastructure, they can better improve efficiency by using an edge compute platform to run custom logic at the edge, and only trickle back the necessary workloads to the cloud, it results in not only a better user experience, but also a reduction in centralized infrastructure costs.

Control and Reliability Are Key

We were initially drawn to Packet for their high performance servers and the level of control that dedicated infrastructure afforded us. However, Packet’s support for both global and local BGP (border gateway protocol) was another draw, enabling Section to use its own IP space on Packet and route traffic reliably to the nearest datacenter on a global basis.

As developers ourselves, we don’t want to own or run any hardware, but Packet allows us to manage physical servers and various “carrier-grade” networking features with the automation we all expect from the public cloud.

This kind of control is even more important as we grow to support ever larger clients on the Section platform. As you might expect, with more demanding use cases we find that our customers require more specialized solutions. The control and performance available with Packet’s bare metal is a powerful tool in our toolbox as we scale.

Endless Edge Opportunities

A current hot topic in the edge computing space is that while many providers are building edge compute infrastructure, there aren’t many clear pathways for developers to understand how the edge works and how to access that edge. That’s the bridge Section crosses. By partnering with companies like Packet, AWS, and Digital Ocean, to name a few, to build out our edge infrastructure, Section provides a programming model that opens the door to the edge for developers.

Packet is holding its annual conference, IFX 2018, November 28-29, 2018 at Bally’s Parking Lot in Las Vegas, where Section CTO Daniel Bartholomew will be delivering a talk titled How Do We Bring Developers to the Edge. Programming (including intimate talks with C-level execs and presentations by leading engineers and partners) will emanate from the core topic areas: plenty of 5G, edge computing, speciality silicon and open source software.