Dynamic Site Acceleration, Application Delivery Network, Front-End Optimization, and Mobile Content Optimization are some of the bread-and-butter protocol optimization services offered by most CDNs. Some are free, and others are available for a hefty price. Over the last decade, CDNs have employed tricks of the trade and made these optimizations a part of their feature set, with the intent of improving page load times. Some of the well known optimizations include domain sharding, keep alive, persistent connections, multiplexing, parallelization, connection pooling, compression, window scaling, and so on. However, CDNs must now consider that sooner or later, Google might take those awesome tricks, and make them a part of their infrastructure. And once that happens, if the optimizations are good enough, Google might push them to become part of a standard.
Google’s innovation in protocol optimization is chipping away at the CDN feature-set. Yes, the Internet is better off as a whole, but the CDN revenue associated with these optimizations will vanish over time. Google’s PageSpeed, SPDY, HTTP/2, and now the UDP-based QUIC are amazing in that they are embedded into HTTP protocol. Although its still early in their life cycle, over time they will mature, and grow in functionality. Google is taking control of these optimizations away from CDNs: multiplexing, compression, persistent connections, and encrypted connections and making it an everyday thing, available to the public free of charge. Conclusion: Google is more interested in content delivery than anytime in the past. Therefore, CDNs must be on high alert, because a killer CDN feature dependent on a protocol optimization, including those in the RAN, might one day become a basic part of an Internet standard, courtesy of Google.