Welcome to a Faster Web

The first commercial deployment of SPDY, a protocol designed by Google to make websites faster, launches today.

SPDY, a protocol Google revealed in late 2009, dramatically speeds up Web page loading by changing the way that browsers communicate with servers. Until now, Google has only tested the research project internally and deployed it on a few of its own sites. But today, the protocol launches as a commercial product.

Website optimization company Strangeloop has built SPDY into its flagship product Site Optimizer, software that sits in between a website and its users, and adjusts the site's code to make pages load more quickly. Strangeloop's customers will have the ability to turn the protocol on easily; in tests, the protocol has sped up websites by 10 to 20 percent.

At first, this will only make a difference for people who visit websites using Google's Chrome browser (the only one that supports SPDY), but Strangeloop expects that it could end up having a big impact on mobile devices as well, since Google is likely to build SPDY into browsers designed for Android.

The speed with which a website responds to users is an increasingly important technical and economic issue. According to the content-delivery network Akamai, people only give a site three seconds, on average, to load before giving up and navigating away. Better performance often means more page views, and thus more interaction with users. For online shopping sites, this translates to higher sales.

It's not just individual sites that have an interest in speeding up the Web. Google has been working to make the Web as a whole faster, reasoning that the entire experience needs to be lightning-quick and smooth to keep people happily using its many services. Google also hopes to entice people to use more Web services and less desktop software (see, for example, its Chrome OS), and the company knows this won't be possible if they struggle with performance.

There are a lot of ways to speed up a website. Changing the protocols that determine how information gets sent over the Internet is potentially the most rewarding but also one of the trickiest. These protocols are fundamental to communication between websites and servers, so they can have far-reaching effects on website performance across all devices. However, to roll SPDY out to the entire world, all browser manufacturers would have to adopt it, and every server would have to support it, says Joshua Bixby, Strangeloop's president. This is a tall order, and so SPDY has "real implementation challenges," he says.

Strangeloop's existing product is, however, well-positioned to help Google make a start. Strangeloop already helps businesses speed up their websites by handling optimization for them without requiring them to change their code or hardware. The Site Optimizer software sits between a company's Web server and the user's browser and adjusts the website's code automatically to make it load faster; this already includes improvements customized for specific browsers. Site Optimizer customers can choose to turn SPDY on, making their servers behave as if the protocol were installed, for customers who visit their sites using Chrome.

Browsers today typically open up lots of connections to a server, in order to start downloading lots of information at once—images, ads, text, and so on. Tom Hughes-Croucher, chief evangelist for Joyent, a company that provides cloud software, explains that, while this does speed things up, the approach also has its problems. Those connections take time to "warm up" and start downloading at their full capacity. Also, they don't prioritize well, so the user might end up waiting for images at the bottom of a page to load when what he's really looking at is on top.

SPDY addresses this problem by opening one connection that is capable of loading many different parts of the page at once. It also allows programmers to manage how pages load, so they can deliver more important pieces first.

Strangeloop's product is designed to handle using SPDY, so customers don't have to worry about writing different code for users who do and don't use SPDY. The company worked extensively with Google engineers to get SPDY deployed and running effectively.

Bundling SPDY with existing optimization products is a good starting strategy for Google, says Eric Hansen, founder and CEO of the website optimization company SiteSpect. He expects Google to eventually include SPDY in its own optimization product, called mod_pagespeed, which is similar to Site Optimizer. Google needs to do whatever it can to get websites to adopt SPDY, he says, because that's the biggest part of its "uphill battle" to gain acceptance for the protocol.

Bixby believes that websites will become more willing to use SPDY when they see its potential benefits. "What's really exciting to me is its capabilities on the mobile side," Bixby says. Google hasn't yet built SPDY into Android's browser, but when it does, the protocol stands to make an even bigger difference. Since mobile Internet browsing is painfully slow and Android handsets have a large portion of smart phone market share, Bixby thinks SPDY could make a difference in that arena. He says, "I would be very surprised if we don't see this in Android in the near future."

Google, Microsoft, and Yahoo Team Up to Advance Semantic Web

A push to add meaning to Web pages to aid search could also enable other kinds of intelligent web apps.

Web of words: This graph of linked phrases lets software understand the meaning of online content. The system is backed by Google, Microsoft and Yahoo. 

Google, Microsoft, and Yahoo have teamed up to encourage Web page operators to make the meaning of their pages understandable to search engines.

The move may finally encourage widespread use of technology that makes online information as comprehensible to computers as it is to humans. If the effort works, the result will be not only better search results, but also a wave of other intelligent apps and services able to understand online information almost as well as we do.

The three big Web companies launched the initiative, known as Schema.org, last week. It defines an interconnected vocabulary of terms that can be added to the HTML markup of a Web page to communicate the meaning of concepts on the page. A location referred to in text could be defined as a courthouse, which Schema.org understands as being a specific type of government building. People and events can also be defined, as can attributes like distance, mass, or duration. This data will allow search engines to better understand how useful a page may be for a given search query—for example, by making it clear that a page is about the headquarters of the U.S. Department of Defense, not five-sided regular shapes.

The move represents a major advance in a campaign initiated in 2001 by Tim Berners-Lee, the inventor of the Web, to enable software to access the meaning of online content—a vision known as the "semantic Web." Although the technology to do so exists, progress has been slow because there have been few reasons for Web page operators to add the extra markup.

Schema.org  may change that, says Dennis McCleod, who works on semantic Web technology at the University of Southern California. By tagging information, Web page owners could improve the position of their site in search results—an  important source of traffic. "This will motivate people to actually add semantic data to their pages," says McCleod. "It's always hard to predict what will be adopted, but generally, unless there's something in it for people, they won't do it. Google, Microsoft, and Yahoo have given people a strong reason."

The Schema.org approach is modeled on one of the more straightforward methods of describing the meaning of a Web page's contents. "The trouble with many of these techniques is, they are really hard to use," says McCleod. "One of the encouraging things about Schema.org is that they are pursuing this at a level that is quite usable, so it is much easier to mark up your website."

If many Web page owners act on Schema.org's suggestions, more than just search will benefit. "This data can be used by any software to cross-correlate things that are related, or to understand the relationship between information from different sources," says McCleod. For example, widespread availability of semantic information might improve artificially intelligent assistants, such as Siri (bought last year by Apple). Or tools able to make good recommendations of, say, news articles because they can know for sure what stories are referring to.

However, the companies behind Schema.org made their move unilaterally, without consulting the World Wide Web consortium (W3C), the standards body for Web technology. "We had no idea this was coming," says Manu Sporny, a member of the W3C's Semantic Web Coordination Group.

Schema.org asks for semantic markup to be written using a format known as microdata, which is not yet a W3C standard, rather than RDFa, a more widely used W3C-approved alternative.

Google has warned that its "crawlers" that roam the Web to build its index could be confused by a page using both microdata and RDFa. Yet Microsoft has previously said its own crawlers have no such problems, says Sporny.

If that confusion isn't straightened out, he says, microdata may become the only standard used at any scale, which would limit the power of the semantic Web, because the alternative can do much more. "RDFa supports use cases that microdata can't—for example, the WHO publishing mortality rates for different countries or adding semantic information to eBook or image files," he says.

Sporny hopes that Google and others behind Schema.org will modify their stance on formats. But he acknowledges that having such large companies embrace the semantic approach is a good thing. "They are saying you will get better results with semantic Web concepts," says Sporny, "and if they encourage more sites to embrace the semantic Web, that will help all kinds of other applications, too."

A Preview of Future Disk Drives

A prototype disk drive based on phase-change memory can outperform an off-the-shelf flash hard disk .



Fast access: This prototype hard drive made using phase-change memory chips can read some data faster than a commercial flash hard disk.

A new type of data storage technology, called phase-change memory, has proven capable of writing some types of data faster than conventional flash based storage. The tests used a hard drive based on prototype phase-change memory chips.

Disks based on solid-state, flash memory chips are increasingly used in computers and servers because they perform faster than conventional magnetic hard drives. The performance of the experimental phase-change disk drive, created by researchers at University of California San Diego, suggests that it won't be long before that technology is able to give computing devices another speed boost.

The prototype created by the researchers is the first to publically benchmark the performance of a phase-change memory chips working in a disk drive. Several semiconductor companies are working on phase-change chips, but they have not released information about storage devices built with them.

"Phase-change chips are not quite ready for prime time, but if the technology continues to develop, this is what [solid state drives] will look like in the next few years," says Steve Swanson, who built the prototype, known as Onyx, with colleagues. It had a data capacity of eight gigabytes and went head-to-head with what Swanson calls a "high-end" 80 GB flash drive made for use in servers.

When it came to writing small chunks of data on the order of kilobytes in size, Onyx was between 70 percent and 120 percent faster than the commercial drive. At the same time, the prototype placed significantly less computational load on the processor of the computer using it. It was also much faster at reading data than the flash drive when accessing blocks of data of any size. The kind of large volume, small read and write patterns that Onyx excelled at are a hallmark of the type of calculations involved in analyzing social networks like those of Twitter, says Swanson. However, Onyx was much slower at writing larger chunks of data than its commercially established competitor.

Onyx was built using prototype phase-change chips made by Micron, a company working to commercialize the technology. The chips store data in a a type of glass, using small bursts of heat to switch sections of the material between two different states, or phases, that represent digital 1s and 0s. In one phase, the atoms of the glass are arranged in an ordered crystal lattice, in the other they have an amorphous, disorganized arrangement.

Onyx's performance springs from the much simpler process of writing data to a phase-change chip compared to a flash chip, which stores data as islands of electric charge on chunks of semiconductor, says Swanson. Flash chips cannot rewrite single bits of information—1s or 0s—on demand. Instead they have to erase data in "pages" of a fixed size and then go back to program in the desired data. That  limits the technology's speed. "It requires a flash memory device to have software keep a little log as it goes along of which data is correct," says Swanson. "With phase-change memory you can just arbitrarily rewrite what you need."

Sudhanva Gurumurthi, who researches computer architecture at Virginia Tech, says the San Diego project is a valuable demonstration of the true capabilities of phase-change memory chips. "Much research has simulated how they would perform, but this gives insights into complexities a simulation can't capture," he says. But it will be the price of the technology that will determine when it becomes a competitive technology, says Gurumurthi.

Gurumurthi's research suggests that using phase-change memory in combination with flash memory could see the new technology reach the market earlier than the day it is cheap enough to be used in dedicated drives. Simulations showed that adding a small buffer of phase-change memory to a flash-based drive could simplify the process of writing small chunks of data, the kind of operation where flash performs least well. "We found it significantly improves performance," says Gurumurthi. "That might be enough to offset the cost of adding a small amount of phase-change memory."

Google Makes Web Pages Load Instantly

The Chrome browser will soon silently fetch pages as you scan search results so that they load without delay.

Google is ready to make waiting for Web pages to load a thing of the past—at least for those pages found using its search engine and the company's Chrome browser. As a user peruses the list of results returned in response to a query, the browser will fetch and load in the background the page it considers most relevant. If a person does click that result, the page will load instantaneously.

The new feature, called Instant Pages, will appear in Chrome and Google's mobile browsers in "coming weeks" but can be experienced today in the developer version of the browser, or later this week in the beta version (the Chrome browser is automatically updated). A video showing the new feature in action has been posted on YouTube.

"We all have broadband, but loading a Web page is still nothing like flipping channels on your TV," said Amit Singhal, Google's head of search ranking, when he announced the new feature today at an event in San Francisco. "We will not be happy until loading a page is as fast as flipping pages in a magazine."

A year ago, Google introduced Instant Search, which loads results live as a person types into the search box. Google studies show that by reducing the time it takes to enter a query, the feature shaves two to five seconds from most searches, said Singhal. He added that Instant Pages could shorten the process of searching for and loading a Web page by a further five seconds.

The company wanted to improve the search experience but realized that speeding up its servers would have negligible effect, Singhal explained. When Google receives a search query, it takes on average about 300 milliseconds to calculate the results and another 400 to send them back to the user's computer. However, it typically takes a person 15 seconds to decide which result to click on and a further five to load a typical Web page.

"In cases where we are confident which result you will select, we are folding the five seconds into the 15 seconds," said Singhal. When Instant Pages algorithms predict with sufficient confidence the most relevant result, they instruct the browser to start downloading and loading that page in the background right away. Users looking at the search results won't notice anything new unless they click the result that was preloaded—in which case it will appear instantly. The feature currently only preloads a single page, said Singhal, although it may include other pages in future.

Other Web browsers could also employ Instant Pages, because Google has released the necessary code for all to use. "We are opening up the code because we want other browsers to implement it—it is good for the users and for the Web," said Singhal.

Other browsers have already implemented a similar feature known as pre-fetching. This, however, downloads just the main body of a Web page, while Google's preloading technology fetches every object on a page and renders it ready for display, even running the JavaScript needed to run sophisticated Web apps, said Peter Linsley, a product manager for Google search.

As the same event, Google announced that users would soon be able to enter search queries on laptops and desktops via speech, a feature previously introduced for use on phones and tablets.

Another forthcoming feature will make it possible to drag an image into the search box instead of entering text. Google's machine-vision technology will analyze the image to find Web pages with images of the same thing. A demo of this feature showed how a decade-old holiday snap could be used to discover the name of the place in Greece where it was taken.