CDN the Elephant in the Room? Strategy
Cisco has noted that “global IP video traffic will be 73 percent of all IP traffic (both business and consumer) by 2017, up from 60 percent in 2012.” ISPs and Content Delivery Network (CDN) operators have already been working hard to accommodate this “video boom.” Now, CIOs have to face the video boom inside their network. With the sudden increase of high quality video delivery, the networks crumble under the load. Regardless of where the traffic has to be delivered, if it is high resolution and high bitrate video, delivering at scale becomes a challenge for both ISPs and enterprise networks.
CDNs have addressed a lot of problems that affect the end user experience. They are able to cache content to offload long-haul connections, they can be closer to the end user to reduce round-trip times thereby increasing overall delivery speed, and they protect the content provider’s origin servers from traffic spikes or excessive load. New features, especially those pioneered by newer players in the space, are making the industry exciting and driving it forward. However, there is a fundamental problem with the current direction: CDN providers operate as closed ecosystems, they are not interconnected and there is no single provider that can deliver with the same quality everywhere, in every single network. The ongoing fierce competition to improve reach, capacity, and features drives the CDN industry and innovation forward. However, if any single CDN provider was to become the single dominant player that would stop development.
In addition, there is a triangle between the CDN provider, the customer, and the end-user with conflicting priorities. From the customer’s perspective, you want to have a single solution for a problem. From an end-user perspective, you want to have the best available quality regardless of how it is delivered or who delivers it. From a CDN provider perspective, you want every single bit of content available for delivery being delivered by your network since volume is most likely how you get paid. That includes delivery to geographies or to foreign networks where other CDN providers are positioned to do a better job for the end-user. What does this all mean? If you are a CDN customer, you have to do serious prioritization between economics, user experience, and integration complexity.
While there is a standard protocol to interconnect autonomous networks (via BGP) and information flow is ensured between remote parties, different CDN providers remain independent and operate as autonomous entities with different (often incompatible) feature sets, capacity, and geographical reach. That means, if you want to receive a “package” (content via CDN) to a certain “address” (your network), the speed and reliability (or even the availability) of delivery will depend on the relationship between the actual CDN and your network. This puts a lot of stress on content providers to select CDNs wisely and pressure on content deliverers to provide quality. The various CDNs have different development directions, competing in delivery efficiency to drive down cost, adding new features, etc. However, they are still primarily interested in locking you in to spend your entire budget with them as opposed to a competitor. In return, they promise they will do anything in their power to improve in all areas, especially in their platform reach and capacity. Is this promise enough? While a single provider is simple to implement and maintain, the risk of vendor failure is constant.
When you choose a CDN provider, you also choose their inherent limitations in network reach and capacity. As a response to this problem, a multi-CDN strategy is common practice in two forms: 1) Limiting the use of CDN features to a level that is supported by all CDNs the same way; 2) Integrating the different features of the selected providers one-by-one to achieve an even field across them, and by that guarantee constant compatibility check and code maintenance for the long-term. The CDNs are interested in delivering more of their client’s content even if it means reduced quality to the customer, because they are primarily paid for bits delivered and not by giving space to another player who may do the job better in a particular region or in a network, for instance. It is up to the company integrating with the CDN to measure and make decisions based on delivery quality for the benefit of their end-users. This industry environment and conflict of interest makes content delivery with consistent quality to all viewers quite a challenge. This is especially true when trying to reach a high number of viewers concurrently in distant networks—and even more difficult when reaching deep into enterprise networks.
To worsen the challenge, the majority of video and application traffic is moving to HTTPS, which renders transparent caching practically impossible. Legacy proxy appliances that provided partial remedy on increased traffic by caching repetitive content are becoming obsolete. While ISPs have been working together with CDNs for many years to resolve this issue, a good solution for corporate networks is yet to come.
There are several ongoing efforts to address this problem, but customer pressure will be also required to drive a change. The work of the Internet Engineering Task Force (IETF) on Content Delivery Networks Interconnection (CDNI) is promising—that could solve the permeability across public CDNs, and maybe even for CDNs on enterprise networks (eCDNs) at some point. The success of the Streaming Video Alliance (SVA) is very important to bring common sense and a manageable future in streaming video, while differences in geopolitical governance in addition to technical compatibility remains a challenge.
In the world of enterprise networks, choosing an eCDN solution is the first of two steps. The second is to integrate the eCDN service with the end user applications and maintain them during their useful lifetime. As SaaS providers have gained momentum, a recent development in the industry is SaaS companies providing virtualized eCDN appliances to help content delivery. With this, SaaS providers can extend their reach deep into enterprise networks and help reduce stress on the network.
Content delivery at scale to both ISP and corporate networks remains a challenge. Companies targeting remote end-users are trying to find the right balance, while corporations look for the solution that can tackle their problem efficiently. There is ample opportunity for existing players to innovate or for new entrants to come into the market and address these challenges in a novel way. As these solutions become available, they will be welcomed by content providers and end users alike as they will bring a better user experience at lower prices and with greater flexibility.