My theory is that within a decade, the world wide web will be left with only porn, spam, and archives of the old way of doing things. The majority of the web will become what Angelfire and Geocities became–outdated and eventually obsolete and removed from the Internet. In the not-so-distant future we will use something different to communicate. There will be no need for .com’s or domains as search and indexing exponentially improves. Simply put, the Internet will be better and faster. Imagine visiting Best Buy or Paul Drecksler’s “online space” as opposed to our websites. What would you call this new web?
Well it turns out, I wasn’t too far off with this theory. Google’s new SPDY protocol is enough for me to say “I told you so!” Whether or not SPDY becomes the new HTTP is questionable; but Google’s “Let’s make the web faster” campaign is on the right track. Let me explain in the easiest way I can.
When the HTTP web protocol was developed, no-one could predict how fast the Internet would grow. The types of applications or webpages built in 1996 are incredibly different than now.
The web runs on TCP/IP and HTTP protocols. Transmission Control Protocol (TCP) allows applications, like websites, to communicate with each other. It was developed in the 1970’s as an improvement to Network Control Protocol (NCP) that was used with DARPAnet, the original Intranet created for the government. Internet Protocol (IP) deals with the routing and delivery of applications. Lost yet? Keep reading it gets simpler.
HTTP’s primary job is to send requests and responses via TCP. This process allows us to visit websites, upload, download, etc. If TCP are train tracks, IP is the train, then HTTP is the ticket booth. It is where we request to visit a place and receive a response on whether or not we can get there.
The downfalls of HTTP are that only one request at a time can be processed. Browsers have adapted to that problem by allowing up to 6 connections per domain at a time but the fundamental obstacle still exists. Also, the client (web surfer) can only request information from the server. The server cannot send the client information even if it knows they need it. The server must wait for a response or request from the user. Both these obstacles lead to a delay in processing and sending data. This delay is called latency.
Wikipedia explains web latency more technically:
Latency is the delay between the initiation of a network transmission by a sender and the receipt of that transmission by a receiver. In a two-way communication, it may be measured as the time from the transmission of a request for a message, to the time when the message is successfully received by the requester.
One of the goals of Google SPDY is to decrease latency by 50%, which means faster page loading. In their controlled testing, Google has been able to reach that goal.
To minimize problems with deploying SPDY, they’ve built the protocol to run on existing TCP. In other words, we’ll still be using the same train and tracks, but the customer service at the ticket booth will increase which makes the whole travel process quicker. The only difference will be that we will need a browser, like Chrome, and a server that runs on SPDY.
Understand that this is just the beginning–the tip of the iceberg. Why stop at improving HTTP? Why not rebuild the entire TCP/IP structure. It will happen in time. Google is smart to work in stages and it makes sense for new technology to work with existing in order for the general masses to affordably adapt.
Google’s SPDY is not the first attempt to revolutionize HTTP. A few examples of previously started projects are SCTP, SST, MUX, and SMUX.
SPDY is open-source and the code will be released by Google in 2010 for developer improvement.
Think about any spy or government movie where the characters are accessing classified government data at lightning speeds using a streamlined interface you’ve never seen before. My first thought was always “Yeah right!” but in another decade, the technology in movies today will seem as outdated as say.. the world wide web.
