Content distribution network (CDN) supplier CDNetworks is seeking a role in the more general computing and 5G market, harnessing edge computing. The company has announced its new Edge Computing Platform Service (ECP), which it says leverages its 1,500 global computing locations, supporting latency of less than 50ms and reducing customers’ operating costs by up to 40%. The CDNs already play in adjacent areas to edge such as video storage and caching.
The service is a container orchestration system built on Kubernetes and Docker for developers to write container-based applications.
“CDNetworks Edge Computing Platform dramatically reduced lag time which resulted in an exceptional online gaming multiplayer user experience,” Mathieu Duperre, CEO and founder of Edgegap, said in a statement. Edgegap is a Canadian start-up focusing on edge and cloud-native computing for online multiplayer gaming.
Rivals such as Akamai and Limelight have signaled their own efforts to sell edge services, while Ericsson is pivoting its former CDN offering into an edge computing play. Limelight has its own online gaming start-up partnership, with Network Next.
Another CDN player is Vecima, a cable industry supplier which acquired Concurrent’s content delivery arm in early 2018. The company’s VP of product and innovation, Kyle Goodwin, in a presentation at the virtual NAB Express event recently, did not hold back in his views of the pitfalls of mainstream CDNs compared to private and hybrid content delivery mechanisms.
Instead of optimizing for single-customer delivery over a managed network, as Akamai or Limelight do, a public cloud CDN like Microsoft 365 must balance demand, which brings bandwidth limitations. Of course, public CDNs do not exclusively deliver video, with these networks handling a whole host of web content and other traffic, meaning the technology cannot exclusively optimize for video the way a private or hybrid CDN can.
This is most critical for live sports latency, because public CDN operators are optimized for throughput and scale, so they cannot also be optimized for latency for a particular event. In other words, public CDNs are not customizable at the cache level, because caches serve multiple different functions.
This is epitomized best by the Netflix model. Netflix realized very early it would be beneficial for itself and the ISPs to cache content closer to users inside ISP networks. This keeps peering points less congested and backhaul less congested, while also allowing for visibility into types of traffic inside a network so ISPs can prioritize this traffic based on their own set of rules, Goodwin explained.
However, peering can cost a lot of money, particularly for ISPs focused on consumers which simply cough up for peering because of the imbalance in upstream and downstream usage, according to Goodwin. Building out high capacity backhaul links is another costly component for ISPs.
If all delivery is IP, operators no longer need to worry about video-specific aspects of delivery in their head ends or in the infrastructure out to customers. That said, components are mostly the same between QAM and IP video but with additional functionality. Formats will vary but the fundamental building blocks are the same.
“Scalability is definitely the hard problem for video delivery. It requires an enormous amount of bandwidth and the throughput of different components can be expensive. Vecima drives complexity down to a level where we can comfortably solve technical issues and strike a balance between scalability and maturity,” Goodwin added.
Vecima’s core message is that a small investment in a little bit of hardware will avoid a large ongoing variable cost to the public CDN. Ultimately, in a software-centric world, Vecima is going against the grain by pushing the idea that a simple piece of hardware is critical in the CDN game – and it makes a compelling case among the plethora of CDN-disrupting alternatives on the market.