Lifesize CTO: tuning software in a video box

Every company is a software company. This central truism pervades throughout every layer of business and, specifically, throughout every layer of the technology industry itself.

Even hardware-centric firms still essentially run as software-centric businesses because (although some specialist equipment is obviously available and no single keyboard, chassis, camera lens, sensor or other unit is necessary of the same build quality) it is the software inside their machines (whether embedded or delivered via cloud services, or both) that provides the differentiating product position that they will ultimately take to market.

Bobby Beckmann is the chief technology officer at Lifesize, a company known for its enterprise-grade high-definition video communication technologies… and one that certainly identifies as a software company.

The Computer Weekly Developer Network (CWDN) spoke to Beckmann to get the lowdown on how the firm puts the software into its total go-to-market proposition.

CWDN: We want to talk software, but let’s get the hardware out of the way first — your firm does offer video conferencing camera systems right?

Beckmann: That’s correct – Lifesize offers meeting room systems that consist of a camera, codec and a Lifesize Phone HD that can jointly handle pretty much any of an organisation’s video and audio collaboration needs. We were the industry pioneers for HD video conferencing systems and we are again with our 4K resolution family of meeting room devices. We also provide ancillary hardware like Lifesize Share that enables wireless screensharing and digital signage in those meeting rooms.

CWDN: Your technology proposition presents hardware that is integrated with your chosen cloud service offering, how have you engineered the provision intelligence in your platform to be able to cope with specific use cases?

Beckmann: All of our hardware is cloud-connected. We have created multiple workflows to allow customers to provision the device to the cloud depending on their specific environment. If they have limited external access, we provide mobile tools to help provision. Once provisioned, the devices can be managed entirely from one central console.

CWDN: What language, platform, tools, methodologies are behind the Lifesize mobile and desktop app and how quickly does it iterate, evolve and change on its own roadmap?

Beckmann: Our web and desktop apps are written in ReactJs. They share common components, especially because our desktop app is running Electron.

Our mobile apps are written in ReactNative, again allowing for significant code reuse between our iOS and Android applications. All of our soft clients leverage a shared common library that is written in Javascript. This library allows us to write business logic that interacts with our cloud service in one place and also lets us leverage it across many platforms. Importantly, we embrace WebRTC at Lifesize and every client uses those libraries to manage our video and audio. We deploy changes to our clients several times a week, whether they are bug fixes or feature releases.

CWDN: You’re big on the whole ‘comprehensive end-to-end encryption (E2EE)’ offering; tell us something about how you work with the code to bring this intelligence to bear in Lifesize.

Beckmann: Our E2EE approach takes advantage of new WebRTC features (Insertable Streams), which allow us to overlay our encryption approach. We did not invent our own means of encryption – that would be a bad idea. Instead, we rely on the browser libraries, which are open source and inspectable. Relatedly, we publish the code for our approach so that it can be evaluated by the broader community. In fact, we have a major reliance on the open source community to create many of the features on our platform.

CWDN: What do you think are the real risks presented by non-secured video conferencing, especially given its surge in popularity in the wake of Covid-19 (Coronavirus).

Beckmann: During Covid-19 (Coronavirus), many organisations have provisioned video conferencing solutions without doing an ounce of due diligence on those providers’ security mechanisms, practices and cultures. The real risk is continuing to blindly use those vendors that don’t seem to have security baked into their DNA and choose to consistently prioritise ease of use and adoption ahead of customer data security and privacy. Once the dust settles a bit, I really do think CIOs, other IT decision makers and line of business leaders are going to have some tough questions to answer from corporate boards about why they deployed the tools they did, what diligence they did or didn’t do and where their organisations’ data and sensitive communications are left exposed.

CWDN: Gartner’s magical analysts think that service providers in the enterprise video solutions market are responsible for systemic-level security of their platform, network and the data it collects from its customers — what’s your approach to data management in this regard?

Beckmann: The first and most important thing we do (or rather don’t do) is that we don’t collect what we don’t need. You won’t find any public tracking services in our app. We also don’t have a way to see what media is being sent during a video call. We do provide our customers the tools to record and store data as they see fit. The only data we need is mostly call data for reporting, but we make it dead simple to delete if a customer requests it. We also make it clear and straightforward to see what data we have collected for a customer.

CWDN: Your firm talks about ability to scale. How have you architected code structures to be able to have that modular ability to expand?

Beckmann: We have tackled this in several ways. As our foundation, we leverage the scalable technologies and services that AWS provides, mainly by building a microservices mesh that is deployed on EC2 instances and Lambdas on the compute side. One of our main design patterns is to build our microservices to scale horizontally by adding more instances depending on load. We accomplish this via a blend of stateless compute and managed allocation of stateful compute. Another core design tenet has been to centralise our business logic and regionalise our media load to best distribute our media services to be close to the customer.

All of this is orchestrated, managed, monitored, developed and deployed with a combination of tools and services including Kubernetes, CircleCI, Github, LogDNA, Serverless, Prometheus, Cloudwatch and Grafana.

CWDN: What makes video industry coders different from, say, fintech programmers or those working in healthtech, transport, core enterprise environments or even game programmers?

Beckmann: The biggest difference for us is that we have to make everything work together. Most software companies can just ship something that everyone downloads and uses in a very controlled environment. If a user complains that their 15-year-old device didn’t work well in a video game, they would be told to upgrade their device. In our world, our goal is to help everyone connect, no matter their hardware or environment, while still clearing a path forward to allow for innovation.

Bobby Beckmann, CTO at Lifesize.

Data Center
Data Management