Transparent caching catching on
A new twist on edge caching looks like it will create a way for multichannel video program distributors (MVPDs) to start making a little money – finally – from carrying over the top (OTT) content.
There are several companies that have recently developed the ability to identify and cache that sort of content on behalf of MVPDs, a process that some are calling transparent caching. They include PeerApp, and the startup Qwilt.
Edge caching is the common on-demand content management technique of taking the most popular content and moving copies from centralized servers and placing them farther out toward the network edge. The maneuver serves several purposes, including getting high traffic items off network backbones while simultaneously creating a higher-quality delivery and viewing experience for those titles.
That process works well for content in libraries under MVPDs’ control. It can even be made to work for content in libraries not under MVPDs’ control, if they know what that content is and can negotiate contractual rights to do edge caching.
For example, it is guaranteed that the top 10 movies released each week to on-demand – available from an MVPD or from an OTT provider – will make up the majority of traffic. By one estimate, 1 percent of all content represents 55 percent of network traffic during peak viewing hours. You just cache those titles.
But what if you can’t get those rights? And what happens when a title not in the Top 10 becomes suddenly popular for whatever reason? Or if content that’s not a feature film comes into demand (e.g. a video like “Gangnam Style” that rapidly goes viral)? Or what if it’s a new software update from a Microsoft or an Apple (which is notorious for doing that sort of thing on short notice)? Or what if it’s content from completely left field? A real-world example offered by PeerApp was people downloading photos from a well-attended wedding posted by the photographer.
The companies recommending transparent caching argue that the OTT content situation is far too unpredictable to simply stick with a top-10 movies strategy.
“Iron Man 3 might flop, and then Gangnam Style comes out of nowhere. If this is a problem of video coming from just one content site, then handling it is easy,” said Dan Sahar, a Qwilt co-founder and its VP of product marketing. “But it’s coming from across all sources. When you don’t know where it’s going to come from, that’s more complicated.”
Netflix and other OTT suppliers propose putting cache servers in MVPDs’ networks, but that approach is problematic. Some of these are proprietary solutions, which means MVPDs could quickly get overwhelmed by a proliferation of cache servers deployed by the increasing number of OTT providers.
The basic idea for these third parties is to cut deals with either the MVPDs or with the CDNs to do transparent caching. The amount of equipment that needs to be added can be minimized. The improvement in quality of service (QoS) ultimately benefits everyone all up and down the value chain.
Qwilt has a software-based solution it is pitching directly to network operators as a means of minimizing the cost of expanding network capacity. It would then turn around and bill CDNs for doing the delivery on their behalf.
One of the attractions of Qwilt’s software-based solution is that it would be implemented by adding only a minimal amount of equipment to already crowded headends and hubs. The alternative, said Sahar, might be racks of equipment to do traffic shaping, deep packet inspection (DPI), and other processes.
Sahar said the Qwilt solution would sit on off-the-shelf hardware (from a supplier such a Dell, IBM, or HP, for example); it would be situated near the CMTS and behave like a router.
The system would monitor traffic through a passive optical tap, probably in a 10G link, Sahar said, make intelligent decisions on what content is popular, and then cache that.
“Iron Man 3 might flop, and then Gangnam Style comes out of nowhere. If this is a problem of video coming from just one content site, then handling it is easy,” Sahar said. “But it’s coming from across all sources. When you don’t know where it’s going to come from, that’s more complicated.”
PeerApp thinks the way to go is straight to the CDNs, who ideally will pay for the solution out of the proceeds of their contracts with their OTT clients, and who will deploy the technology. The deployment will be on the MVPDs’ premises, but the benefit to MVPDs should be so cut-and-dried that they’ll encourage the implementation of the solution. Why?
- CDNs get credit with their OTT clients for having negotiated better QoS on their behalf;
- MVPDs get credit with their customers for providing better QoS on OTT services, and they get paid to install the solutions on behalf of the CDNs (and, to belabor the point, that means no capex responsibility for the solution);
- The MVPDs’ subscribers get better QoS on OTT content
- The OTT provider gets better QoS.
PeerApp it has several hundred installations around the world, 30 percent to 40 percent of them with mobile operators, but has only developed the transparent caching model – it refers to it as “Content Service Extension” – in May. PeerApp director of marketing Josh Adelson said the company is testing that model in trials right now with a large Asian provider and a European operator he declined to name.
Concurrent meanwhile recently announced having been awarded a patent for demand-based edge caching of video content.
The patent was awarded for a method and system of prioritizing and intelligently caching video content at the edge of a content delivery network based on asset demand. By caching popular video content at the edge, more requests can be fulfilled from within the local network and demands placed on regional and metro networks can be dramatically reduced.
Concurrent Computer CTO Bob Chism said “Intelligent edge caching is an important element of any content delivery network. With video accounting for a growing proportion of the traffic being managed through CDNs every day, our customers are interested in solutions that can reduce the amount of core network bandwidth consumed by video streaming applications. Our caching technology ensures content is stored and delivered as efficiently as possible, minimizing the use of our customers’ valuable network resources.”