The following are ten servers that have enabled IT world-beaters to develop technological advancements that transformed the way we work and live. While the people behind the systems are the real brains, this list highlights the hardware they relied on.
Without further ado, in no particular order, here’s our list of the Ten Servers that Changed the World.
The Sun Ultra 2 may seem like an unlikely candidate to make this list, but it steps up as the server which first hosted Larry Page and Sergey Brin’s Backrub search engine – which, of course, eventually evolved into Google.
In 1998, Backrub was hosted on a Sun Ultra 2 with dual 200Mhz CPUs and 256MB of RAM at Stanford University. The famous image of the computer case partially made up of legos (pictured) isn’t actually the Backrub server, but rather its enclosure for external storage. (There were also a couple of Intel Servers and an IBM RS/6000 F50 in their network.)
This is quite a humble beginning, considering there are now over 450,000 servers in Google’s datacenters around the world. The simplicity of its search engine and its relative results blew away their competitors. (And it all started on an Ultra 2.)
NeXT and its NeXTCube are often cited as infamous flops. NeXT was a bold company, led by Steve Jobs during his Apple sabbatical, that didn’t quite live up to expectations. Despite its shortcomings, the NeXTCube will always have a place in history as the very first web server.
The World Wide Web was born on a NeXTCube with a 25Mhz CPU, 2GB of disk and a gray scale monitor. Sir Tim Berners-Lee put the first web page online on August 6, 1991 while working for CERN in Geneva Switzerland. He designed the first web browser and editor, WorldWideWeb, on the NeXTSTEP OS. Berners-Lee continues to shape the web world as the founder of the W3C (World Wide Web Consortium), a researcher at MIT and recently as an advocate for the protection of Net Neutrality.
In 1996, Apple Computer acquired NeXT – and several components of the NeXTStep OS would be crucial in the development of Mac OS X. Sun Microsystems had also made investments in NeXT and ported some of the OS’s components into the PA-RISC SPARC systems. Incidentally, the NeXTCube was also used by John Carmack to develop the games Wolfenstein 3D and Doom.
While Alexander Graham Bell’s first telephone call was clearly documented as “Mr. Watson, come here,” the world will never know what, exactly, was transmitted between two side-by-side DEC PDP-10 nodes in 1971.
The legendary Ray Tomlinson of BBN, sent the first email over these nodes via the ARPANET network, but contends that he doesn’t recall the characters he first transmitted, stating that it was “something like ‘QWERTYUIOP.” Nonetheless, network email was born, and just over two decades later became the foundation for electronic communication, breaking down barriers and flattening the world. We can also thank Ray for bringing the “@” symbol into our daily use, as he decided to assign it as the unique character within email addresses.
Unlike the PDP-10 pictured here, Tomlinson’s PDP-10s did not have any type of monitor or green screen, but rather would output to a printer reel. The PDP-10 was a great success for DEC and was eventually used by companies like Microsoft, who developed several versions of the BASIC language on it. The model was widely adopted by universities, and in fact, even Bill Gates learned on one in college. And Gates’ counterpart, Paul Allen, seems to have a place in his heart for them as well – owning a working model in his personal collection, documented online at PDP-Planet. The CGI for the movie TRON was rendered on a PDP-10, too.
Those interested in testing their programs in 36-bit goodness can find several PDP emulators on the net to play with.
Before the internet became “a series of tubes,” SAGE, the first fully operational wide-scale network, actually was.
SAGE was designed by IBM at MIT in 1956 for the AirForce. It was based on several of the IBM AN/FSQ-7 Intercept computers, and performed as an air defense system. Each AN/FSQ-7 used 55,000 vacuum tubes and occupied almost a 1/2 acre of datacenter space. It was the biggest computer in history and its size will most likely never be surpassed.
The AN/FSQ-7 Intercept was a 32-bit dual processor system with hot-pluggable power supplies, a modem and sold for $238 million. It turns out that SAGE probably wouldn’t have worked for its intended purpose of air defense, but the AN/FSQ-7s stayed in production until at least 1985. They served well for air traffic control and were also a popular backdrop for Hollywood command centers.
When world champion chess player Garry Kasparov lost a chess match to IBM’s Deep Blue computer on February 10, 2006, the world was at attention.
Deep Blue ran on the AIX OS and was built on a 32-node RS/6000 SP RISC system. It could generate 200 positions per second and rank the “goodness” of each one. It didn’t necessarily create a new technology or make significant advances towards one… so how did it make this list?
Over 100 years, earlier the industrial revolution had begun to make manual labor more efficient and reduce opportunities for man. Fears manifested in fables such as that of John Henry, who represented the best laborer man could offer versus machine. Now in the 1990s, automation seems not only a threat to hard labor, but also to those who use their brains instead of their bodies.. I believe Deep Blue versus Kasparov became more than a marketing event for IBM; it turned into a modern day fable representing our collective fears of what technology could accomplish, and therefore what it could take away. Of course, this fable is exaggerated because new opportunities will inevitably arise with new technological developments.
As an interesting side note, there are theories that Deep Blue may have had some help. Kasparov believes the machine did not act appropriately, and other research has shown intriguing evidence as such. IBM denies any interference.
In 1965, two servers on opposite coasts were networked together, driving home the golden spike in a transcontinental wide area network.
Thomas Marill came up with a strategy to connect distant computers and transfer data across telephone wires. Marill then hooked up with Larry Roberts and ARPA to make it happen.
The Lincoln TX-2 at the Lincoln labs in MIT, designed by Wesley Clark, was connected with an IBM Q-32 (AN/FSQ-32) in Santa Monica, California at SDC (System Development Corporation) Headquarters. In 1966, Marill and Roberts documented their experiment and co-wrote Tward a Cooperative Network of Time-Shared Computers.
When the first LISTSERV was created in 1981, the doors opened up to group email collaboration. (Not to mention list spam, off-topic discussions and flame wars.)
The original LISTSERV was hosted on an IBM VM mainframe over BITNET (Because It’s Time NETwork). BITNET would later incorporate DEC VAX systems into its network as well. Ira H. Fuchs of CUNY and Greydon Freeman of Yale decided to connect their universities using a leased telephone circuit between their mainframes.
By 1982, BITNET reached across the US and into Europe, creating a worldwide network. The network peaked in connecting over 1400 organizations in 49 countries, but would sharply decline from here on out due to the growth of the internet.
The DEC PDP-7 was released in 1965, but it was in 1969 when Ken Thompson of Bell Labs and his team would develop the Unix OS. He would have liked to have gotten his hands on a PDP-10 or an SDS Sigma 7, but funding was refused, so the PDP-7 had to suffice.
In brief, Thompson was familiar with the MULTICS OS and had been developing a game called Space Travel on it. He wanted to develop advanced functions such as rotating planets that didn’t seem possible in the current iteration of MULTICS. He was inspired to come up with a new OS that could be programmed on the PDP-7 and his team dubbed it UNICS (an emasculated MULTICS). The name obviously evolved into “Unix,” as did the OS itself once it was developed on more advanced systems such as the PDP-11.
Due to the fact that UNIX was developed on the PDP-7 and its printer reel output (with no monitors or terminals), it still remains true that UNIX is composed of very sparse commands and responses.
We could have gone with the XBOX 360 here, due to the number of people who are hacking Linux onto it, but the fact that the Playstation 3 is going to support and distribute Linux gave it the edge. There are also programs already shaping up to use the PS3 as a wide scale distributed computing system.
Although it is yet to be released to the general public, the PS3 looks like it has the potential to put server power in the hands of thousands who havenÃ¢â‚¬â„¢t had it before. The distributed computing options will also supply additional processor nodes to those networks that need all they can get, such as SETI or Folding@home.
The original users of those systems above, such as the 1/2 acre SAGE system, must be blown away by the processing power that is packed into this home console such as…
– 3.2Ghz Cell Broadband Engine CPU
– 60GB ATA Hard Drive
– 256MB RAM
– 550Mhz RSX Graphics Processing Unit
– Built-in Network Capabilities