Constructing the Markley Network Fabric

By R. Leigh Hennig | February 28, 2022

One of the most common questions I get asked is something along the lines of, ‘how do we architect our network?’ Customers want to know what kind of devices they need in their network, what capabilities and speeds they should support, how do they ensure redundancy. It’s a loaded question, and my initial answer is always the same: it depends. It’s a great conversation starter, and one of the most interesting and enjoyable parts of my work.

The reason it depends is because to design a network to meet current and future needs, you must understand the network profile. What kind of traffic is going to be sent across the wire? How many users are there, and what are their expectations? How much data do they need to send at peak? What are their latency needs? Usually, the answers to these questions have a familiar theme. They want to:

- Send a lot of data (hundreds of gigabytes at a minimum, and often hundreds of terabytes or more)

- Send the data quickly, with as low latency as possible (anything over an RTT, or round-trip time, of 10ms may not be acceptable, depending on the application)

- Secure their data in transit, not just at rest

- Ensure reliability, protecting against hardware or link failure

All of this must be done without over designing the network, and it must be done frugally. The network must be reliable, easy to operate and maintain, and be cost-effective.

Even though many of these considerations are common amongst network operators, certain organizations have network and application profiles that make their requirements more bespoke. This is particularly applicable to pharmaceutical and biotech companies; environments where Markley has extensive experience and is especially adept at working with.

The Markley Network Fabric was born from a mission-critical request from a global pharmaceutical company responding to the COVID-19 pandemic. The request was simple: build us a network that connects our researchers, laboratories, and manufacturing facilities to unblock their efforts to save lives. This network needed to have the lowest latency possible, support vast amounts of data transfer from scientific instruments like sequencers and imagers, facilitate direct connectivity to public cloud services like AWS, Microsoft Azure, and Google Cloud, and all but guarantee zero downtime. No pressure, right?

It didn’t take us long to learn from dozens of other biotech companies (the Boston Metro area is rich with biotech and pharma, being home to over 1300 of them) that they all had the same demands. An additional problem presented itself by way of the disconnected and diverse locations where these companies were based. It wasn’t enough to tell these organizations that they needed to establish a presence in our data center—the Markley Network Fabric had to be designed in such a way that made it flexible, modular, and deployable to any location, regardless of locality. A “lift and shift,” as the processes of moving your existing equipment to a new datacenter is known, was simply not feasible for many.

We succeeded in that request. While Markley has a breadth of experience working with leading organizations from diverse industries such as financial services, security, manufacturing, and more, we are experienced experts with biotech and pharmaceuticals. We’ve spent a lot of time listening to their needs, learning their network profiles, solving their unique problems, and building the Markley Network Fabric according to their requirements and feature requests.