GCN LAB IMPRESSIONS
Why Google's SPDY is none too swift
Good old HTTP has been around a while. It’s been like a quiet warrior, a protocol that enabled the Internet to function the way it does today, yet without too many people really thinking much about it. In brief, the Hypertext Transfer Protocol makes sure that all those bits and bytes get from servers to your computer in one piece and in the right order, not that anyone notices.
Nobody even puts http:// at the front of URLs anymore. Browsers insert this automatically for anyone who types in a Web address. But it, and its more secure HTTP Secure cousin, are there on almost every Web address in the world.
I think I first encountered HTTP in 1991, close to when it was released. It was still called Version .9 back then. It didn’t get an upgrade to Version 1.0 until 1996 and was tweaked again in 1999 to Version 1.1. Not a lot has been done with it since then. It works and seems to do well, and people are not really into fixing things that aren’t broken.
Except Google of course. Google doesn’t think HTTP is such a hot property. It is working on a new protocol called SPDY (short for speedy) that the company says should replace the venerable HTTP.
90 percent of secure HTTPS sites are vulnerable, study finds
Google says the advantages of SPDY over HTTP are that it uses compression, multiplexing and prioritization technologies to speed packets along. For example, each SPDY-enabled client keeps track of which headers it’s sent in previous requests and doesn’t resend them. If it does have to resend a header, it does so compressed to the SPDY-enabled server. Google has graphs and charts to prove all this, of course, and some companies such as Twitter have implemented the ability to use SPDY if it’s available on both ends of a transaction.
Google is trying to make this a standard protocol, and the group working on HTTP 2.0 is said to be considering implementing SPDY into its build. Because SPDY works on the back end, users probably won’t notice much of a difference, other than increased speeds.
On the one hand, I think speeding up the Internet by trimming some of the fat in transactions is a good thing. On the other, I don’t like how Google owns the SPDY protocol. And the company has not stated that it wants to make it freely available to everyone.
I also know that Microsoft is also working on its own version of a faster HTTP protocol. These titans never work too well together, and I wonder if this might lead to a new format war. I still remember how Internet pages looked different when viewed in Netscape and Explorer, and even today, they are not completely universal on every browser, as was intended when the Web was created.
I wouldn’t be so worried about SPDY or other experimental protocols if companies had historically worked together to make the Internet a better place. But that has almost never been the case, as my colleague Greg Crowe pointed out earlier this week when he wrote about the latest wireless format wars.
I’m all for improving HTTP, but it’s too important to mess with willy-nilly, or to open up a bunch of competing standards so that the Web ceases to be universal. Let’s hope that doesn’t happen, but I won’t hold my breath.