S&M vs. SPDY: Microsoft and Google battle over the future of HTTP 2.0

HTTP, the protocol that underpins almost every inch of the world wide web, is about to make the jump from version 1.1 to 2.0 after some 13 years of stagnation. For a long time it looked like Google’s experimental SPDY protocol would be the only viable option for the Internet Engineering Task Force (IETF) to ratify as HTTP 2.0, but now out of left field comes a competing proposal from Microsoft.

Lumbered with the truly awful name of HTTP Speed+Mobility, or HTTP S&M for short, Microsoft’s vision of HTTP 2.0 is mostly very similar to SPDY (and it admits as much on the Microsoft Interoperability blog), but with additional features that cater towards apps and mobile devices. “The HTTP Speed+Mobility proposal starts from both the Google SPDY protocol and the work the industry has done around WebSockets,” says Jean Paoli from the Microsoft Interoperability team. WebSockets in this case refers to a feature in HTML5 that allows websites (or indeed web apps) to open up bidirectional, real-time channels with remote servers over TCP, which is something that HTTP (nor SPDY) is capable of.

What’s so special about SPDY, then, and why is Microsoft (uncharacteristically?) keen to build HTTP S&M on top of it? In short, SPDY (short for “speedy”), is incredibly quick. Basically, with HTTP 1.1 a computer can only request one resource at a time over an HTTP/TCP connection, and each of those requests needs to be sent across the web, acknowledged, and dealt with in sequential order. With SPDY, multiplexing is used to squirt unlimited numbers of HTTP requests over a single TCP connection, and each request can be given a priority (some parts of a website, such as the HTML and CSS, might need to be loaded before the images, for example). Furthermore, SPDY uses TLS (transport layer security) and compression throughout.

In testing, SPDY loads web pages about 40% faster than HTTP. In practice, SPDY is currently supported by Chrome (though it will be on by default in Firefox 13), but only a handful of websites serve SPDY content (mostly Google’s own services). If you’ve ever wondered why Chrome seems to load Google search or Gmail so much faster than other browsers, look no further than SPDY.

In short, the entire purpose of SPDY is to speed up the web — which isn’t a bad thing, and nor is it surprising, considering Google’s fanatical penchant for speed, but Microsoft is basically saying thatspeed isn’t everything. With HTTP Speed+Mobility, Microsoft is saying that we should also take into account factors such as battery life and bandwidth cost, both of which will play a big part in Windows 8 in specific and mobile computing in general.

To this end, HTTP S&M wants to do away with SPDY’s blanket use of encryption and compression, both of which require a lot of CPU cycles and therefore would be no good for an almost-dead mobile phone or low-power sensors. Likewise, server push, which is a standard feature in SPDY, would be dangerous for devices on a metered connection. In their place, Microsoft proposes that these features are added through extensions, rather than making them built into the HTTP 2.0 spec.

In essence, Microsoft wants HTTP 2.0 to be as flexible and long-lived as HTTP 1.0 and 1.1, rather than the brute force, one-size-fits-all approach that Google seems to be championing. Considering just how fast the internet moves, I think Microsoft’s approach is probably wiser.

Read the HTTP Speed+Mobility IETF draft, or read more about SPDY

FCC fires FUD at the idea of a UN-controlled internet

In a recent editorial at The Wall Street Journal, FCC Commissioner Robert McDowell blasted the upcoming ITU World Conference on International Telecommunications (WCIT-12). According to McDowell, Russia, China, and their allies at the ITU want to monitor all internet communications, allow foreign companies to charge for international internet traffic “perhaps even on a per-click basis,” impose economic regulations, take over ICANN (Internet Corporation for Assigned Names and Numbers), and conquer the Internet Engineering Task Force.

McDowell reaches a bombastic crescendoby claiming that the treaty will more-or-less destroy everything, everywhere, writing: “Productivity, rising living standards and the spread of freedom everywhere, but especially in the developing world, would grind to a halt as engineering and business decisions become politically paralyzed within a global regulatory body.”

The FCC Commissioner’s threat assessment is completely out-of-step with the US government’s opinion, as shown in a leaked memo from January 23, 2012. The memo notes that while there was “great and widespread concern” a year ago that WCIT-12 would be a battle over the role the ITU should play in internet governance, the US spent 12 months working to limit the scope and nature of the issues that will be considered at the treaty negotiations. As a result, “There are no pending proposals to invest the ITU with ICANN-like Internet governance authority. Neither cybersecurity nor Internet governance predominate discussion in any region.”

Among the charges leveled at the ITU are claims that the treaty could “Impose unprecedented economic regulations such as mandates for rates, terms and conditions for currently unregulated traffic-swapping agreements known as ‘peering.’” As we’ve said, there’s literally no such agreement under consideration — but the inclusion of this point sheds light on why certain parties are so interested in keeping this issue in the news.

Under the current unregulated peering system, foreign ISPs pay US ISPs a fee to carry internet traffic, which means US companies make a tidy sum of cash off foreign access. If internet servers were truly decentralized — the “Balkanization” McDowell fears — US ISPs would end up paying considerably more money to their foreign counterparts.

Those bright white lines aren’t just revenue sources, they’re control linkages. If you work for the MPAA/RIAA or back laws like SOPA and PIPA, those links are absolutely vital. Any attempt to create an international system of internet governance would weaken the RIAA and MPAA’s efforts to implement SOPA-style censorship. Both bills were aimed at restricting and controlling foreign internet traffic, which means both intrinsically assumed that such traffic would be flowing through the United States.

An equally distributed intra-planetary internet would still take geolocation into account for routing and access purposes, but would effectively eliminate the concept of “foreign” websites. SOPA and PIPA were meant to be palatable to the general US population precisely because they exploited an us/them mentality and claimed to be protecting America. If internet control were to shift towards nations that favored fewer copyright restrictions, internet access as a human right, and limited punishment for piracy, it would be a serious threat to content distributors.

McDowell’s claims are factually inaccurate and hyperbolic. They paint a false dichotomy between the idea that the internet today is a free-wheeling, uncontrolled frontier, while the alternative is a fascist state. The internet, as it exists today, is highly regulated. Some of that regulation was inherited or expanded from the old laws governing telephone access and line-sharing, some of it is applied via laws like the DMCA. ICANN is not a direct arm of the US government, but it’s a far cry from a private corporation. The publicized debates around net neutrality and the FCC last year are further evidence that the idea of an unregulated internet is a fallacy.

At the other end of the equation, no one advocates handing over complete control of the internet to the likes of Russia, China, Myanmar, and Iran. There’s no reason not to open internet governance slowly and gradually, unless you represent a faction who views such a process as an unacceptable loss of control. Regardless of how you feel about the issue, McDowell’s editorial only clouds the debate with demagoguery. It’s a blatant attempt to fire people up emotionally with virtually no grounding in objective fact. The internet is going nowhere, regardless of what happens at the upcoming meeting. Ultimately, however, this isn’t a debate about whether the internet is regulated, but an argument over who should control the regulatory process. If US lawmakers continue pushing bills like SOPA and PIPA, they may find an increasing number of US citizens who think the UN is a more attractive alternative — a concept editorials like this are meant to thwart.

Source