The Second World War demonstrated the devastation that could be caused by even conventional bombing…and was capped by the nuclear destruction of Hiroshima and Nagasaki. With the intensification of the Cold War and the first Soviet atomic bomb test…and the Communist aggressiveness demonstrated by the outbreak of war in Korea…air defense of the United States became an issue of very high priority.
During World War II, the British had been successful with their innovative network of radar stations linked to command centers at which the positions of friendly and enemy aircraft were plotted continuously and orders issued to fighter squadrons and antiaircraft gun sites. In the postwar era, though, the increased speeds of combat aircraft, combined with the utter devastation that could result from a single failed intercept–one plane, one bomb, one city–drove the view that something better than manual plotting would be required.
Although digital computers were still very much in their infancy in 1953, the solution to the air defense problem chosen in that year was a computer-based system to be known as SAGE…the Semi-Automatic Ground Environment. Real-time information from multiple radar sites flowed in digital form to the computers at the SAGE Direction Centers. The computers tracked the targets, friendly, unknown, and enemy, and displayed them on dozens of video displays at each Center. Battle-management personnel at these displays made the determination of which enemy targets should be engaged with what priority, and what friendly aircraft should engage them, and the computers then calculated the optimum intercept courses. For certain fighter aircraft types, the interception commands could be relayed directly via datalink, obviating the necessity for voice communication. SAGE Direction Centers also had control over high-speed BOMARC antiaircraft missiles…these carried small nuclear weapons intended to ensure that a near miss would not allow enemy bombers to escape.
At the heart of each Direction Center was a pair of computers, AN/FSQ-7, duplexed for reliability. Each pair contained fifty thousand vacuum tubes, covered almost an acre of floor space, and consumed about 3 megawatts of power. (Some sources cite the 50,000-tube number as being for each computer of the pair–either way, it’s a LOT of vacuum tubes.) Here’s a fairly well-done recent article about the SAGE project. Note, however, the author’s comment about “thousands of people all over North America constantly scanning their radar screens for Soviet attacks, all hankering for an opportunity to launch a radio-controlled nuke.” I wonder: does this guy really believe that the airmen at the SAGE scopes were really looking forward to a nuclear war, or did he just think that’s the sort of thing that would play well with his editors and his audience?
Developing the hardware required for SAGE was a challenge; developing the software even more so. IBM’s Tom Watson Jr explained the issue: “In those days computing was typically done in what was called batch mode. This meant that you would collect your data first, feed it into the machine second, then sit back for a little while until the answer came out. You could think of the batch processor as a high diver at a circus–each performance involves a lengthy drum roll in preparation, a very fast dive, and then a splash. But the SAGE system was supposed to keep track of a large air defense picture that was changing every instant. That meant it had to take a constant stream of new radar information and digest it continually in what is called “real time.” So a SAGE computer was more like a juggler who has to keep a half dozen balls in the air, constantly throwing aside old balls as his assistant toss him new ones from every direction.”
Each SAGE computer had to accept continuous streams of data sent from the radar sites over leased telephone lines (this was the first large-scale example of computer networking) as well as flight plan data from the airlines and the FAA, perform tracking and intercept calculations, manage the display of information on dozens of screens, and–when required–transmit data link commands to properly-equipped interceptors and direct the flight of one or more BOMARC missiles. All of this had to be accomplished with a computer rated at only 75,000 instructions per second, far slower than whatever device you are using to read this post.
The initial operational deployments of SAGE were in 1958, and the full nationwide system was in place by 1962. This was a vast project, requiring not only the production and development of the computer systems and interfaces, but also massive construction efforts. Total cost was somewhere around $8 billion, which translates into roughly $63B in 2013 dollars. SAGE ran until 1984, when it was replaced with more modern technology.
Some have asserted that SAGE became obsolete when the Soviets launched Sputnik in October 1958, but I think this is a considerable overstatement. Development and deployment of Soviet intercontinental missiles could not take place instantaneously, and certainly at the time of the Cuban Missile Crisis, in 1962, manned bombers were still an important factor in the balance of forces calculations of the two powers. (Also, of course, SAGE technology was–at least at the conceptual level if not the hardware and code level–directly relevant to the development of missile warning systems.)
Whether or not SAGE was a wise investment from a national defense standpoint, it certainly had a major technological impact. At the hardware level, it was the first major system to use magnetic core memory. Experience gained in on-line systems development led directly to the use of computers for airline reservations systems and air traffic control–indeed, one of the SAGE Direction Centers was actually used for ATC purposes for a while. SAGE was the first large-scale system to use visual displays, and the use of the light gun to select objects on these screens was the predecessor of the mouse. SAGE pioneered to development of data communications, which would lead eventually to the Internet.
Here’s another recent article on SAGE. It’s mostly political ax-grinding, and I don’t think it’s very accurate; I’m linking it mainly because the discussion thread includes comments from people who actually worked on the system. (Another recent article, the link to which I have lost, referred to SAGE as being motivated by the “Red Scare,” as if it had been irrational to be concerned about nuclear weapons in the hands of Josef Stalin and his successors.)
While researching this post, I encountered several interesting items of SAGE lore:
–Early in the operational deployment of SAGE, someone (of course!) figured out how to use the system to create an image of a pin-up girl on the display screen. Here she is. (A comment here notes that displaying pictures of unclothed young women is somewhat unrelated to national defense..unless one takes a very long view.)
–Each SAGE Direction Center was connected, via redundant land lines, to the BOMARC missile sites over which it had control. If one line failed, a black box at the missile site would switch the listening watch to the other line. An analyst discovered that if this second line also failed, the equipment would just keep listening to the noise on the bad line. He did a little math and showed that in this situation, random noise would generate something that looked like a FIRE command, probably within two minutes…receipt of this command would cause the missile to erect to a firing position and begin its launch sequence.
This actually happened, only two weeks after the analyst wrote his report, scaring the hell out of the personnel at a BOMARC site near Washington DC. Apparently there was actually not much danger of an actual inadvertent launch: a whole sequence of acceptable guidance commands would have to have been received during a short window, and the analysis showed this to be very improbable. Still, very spooky for all involved, I imagine.
–SAGE makes an appearance in John Updike’s 2007 novel Villages, which I have not yet read.
I was working at Douglas in 1958 and programming an IBM 650. The main plant, across the street, had 704s which was a much bigger and more powerful mainframe. While I was there, the 704 was replaced with the 7040, which was core memory instead of vacuum tubes. I work a couple of times a week at a nearby military recruit center and drive by the old plant site. There is no remnant of the old plant or the wind tunnel building where I worked 55 years ago.
I spent 16 years of my 22-year USAF career as a radar maintenance type in the SAGE system… from 1964 until 1980… serving in places like Lompoc AFS, CA (think Vandenberg AFB), Keno AFS, OR; North Bend AFS, OR; Boron AFS, CA; and Fortuna AFS, ND. There was a point in time when every square inch of American soil in the Lower 48 (and most of Alaska and Hawaii, too) was swept by a radar beam every 12 seconds. All those radar sites and the Direction Centers they fed are mostly gone now but you can get a flavor of them and the people that manned the sites at Radomes.org.
Very interesting article, thanks. I am amazed at how big the computers were and how much power was needed to make them run. I also did not know about the light gun.
I was in Army Air Defense in the early 70s and learned all about SAGE – although by that time it was mainly history.
And SAGE was linked to the Nike-Hercules 2 stage solid fuel missiles.
People these days don’t appreciate the reliability of computers but in the tube days you had a lot of downtime – and them someone would have to find the bad tube(s).
My Dad tells me that Irving Reed sketched out the SAGE network during a meeting lunch break. I don’t know how accurate that story is, but Irving Reed was a bright guy and there is probably some element of truth in it.
Our IBM 650 had a big room, about 20 by 30, with a raised floor for the cables and the air conditioning kept the room temp at about 65. Our desks were outside as it was too cold to stay in there.
When I was a freshman in college, about 1956, I later learned there was a big crisis with the Soviets when they developed a new radar system and first turned it on. They discovered US planes, mostly B 47s, that were flying over their territory and they had been unaware of it as their radar had not been able to pick them up’
I knew about the U2 in 1958 from gossip at Douglas. My future father-in-law and I discussed it when I came over to pick her up for our first date. We never did go out that night. He and I spent the evening telling stories and she and her mother made dinner.
@Mike K – you & I remember that era – where you would go into a “clean room” of raised white flooring (usually white), underneath which ran cables and coolant pipes for the CPU – it was always a very impressive sight. The sound of the low and powerful hum of cooling fans permeated the room.
With the advent of the microprocessor, that stuff started going away but interestingly the fastest micro CPUs now are using liquid cooling, although on a much smaller scale ;-)
This is an interesting post. And in my section of the Army, we had a symbiotic relationship with the Air Force. We usually worked under an Air Force squadron. The Air Force didn’t want any misunderstandings when it came to a decision to fire a missile at a plane.
But Buck’s post reminded me that SAGE was a lot more than ground to air defense. It fed information to interceptors.
Through all of this, transponder information was critical – Our planes had to have a proper numeric sequence “squawked” or they would be a target.
Incidentally the term “squawked” – where you have a sequence of numbers on a set of dials – being fed back with the radar’s returned “bounce” – had its origins in WW2. It’s a term that is still used today for civilian aviation.
Was trying to find the origins of the term in WW2 but the “all knowing oracle” on the Net has come up dry. But I know it is of WW2 origins.
From Wikipedia:
The aviation transponder was originally developed during World War II by the British and American military as an “Identification friend or foe” (IFF) system to differentiate friendly from enemy aircraft on radar. The concept became a core of NORAD technology in the defence of North America during the Cold War.[citation needed]
This concept was adapted in the 1950s by civil air traffic control using secondary surveillance radar (beacon radar) systems to provide traffic services for general aviation and commercial aviation.
Secondary Surveillance Radar
Main article: Secondary surveillance radar
Secondary Surveillance Radar (SSR) is referred to as “secondary”, to distinguish it from the “primary radar” that works by passively bouncing a radio signal off the skin of the aircraft. Primary radar determines range and bearing to a target with reasonably high fidelity, but it cannot determine target elevation (altitude) reliably except for at close range. SSR uses an active transponder (beacon) to transmit a response to an interrogation by a secondary radar. This response most often includes the aircraft’s pressure altitude and a 4-digit octal identifier.[7][8] </i?
On the origin of the term “squawk”…apparently in early days, you could actually hear the transponder replies on the aircraft’s communications radio…and it sounded like a parrot squawking.
The ADS-B system for air traffic control, which is gradually being implemented, is based on aircraft continually and automatically reporting their positions, which they know from onboard GPS, via transponders (though I guess the term doesn’t really fit anymore since since they’re now not responding but rather initiating.)
Data centers these days are getting even bigger than in the vacuum-tube mainframe era…see, for example, Rackspace’s new facility in the UK, which is on a 12-acre site and will draw 10 megawatts.
Meant to include a video with the post…here’s an IBM ad featuring SAGE from sometime in the 60s:
http://www.theatlantic.com/video/index/251592/ibm-sage-defense-1960s/
I always thought the BOMARC was one cool looking missile.
BOMARC: 200-mile range, mach 2.5+, guidance by datalink from SAGE until it got close enough to the target for its own radar to take over.
Main propulsion was a ramjet, but the missile had first to be accelerated to a speed at which the ramjet could start up. Reading the Wikipedia article, I see that the initial acceleration phase was done by a liquid-fueled rocket engine until about 1960, when it was replaced by a solid-fuel rocket. The liquid fuel was very dangerous and corrosive stuff, so the missile could not be fueled until immediately before launch. So if the event I mentioned when a false LAUNCH command was received happened when the missiles were still liquid-fueled, then there would have really been no danger of an accidental launch…still would have been a scary event, though, because a genuine launch command would only have been received if it were believed that the US were under nuclear attack.
I think there are a lot of empty SAGE centers around the country – lik Titan missile silos
There are lots of old Nike sites scattered around the country too. A lot of what are now suburbs have addresses containing the words “Nike Base Road”
There was one on “the point” on Lake Michigan at 55th St. near the U. of Chicago.
There are at least a couple in South Florida, one adjacent to Card Sound Rd. near Key Largo that is a ruin. Another, in Everglades National Park, was recently restored with a Nike missile on static display and is open for tours.
There’s a former Nike base just across the Golden Gate Bridge in what is now the Golden Gate National Seashore.
One of my sons was busted for trespassing at a teenager party there – he had to go to Federal court!
Lots of Nike information here:
http://ed-thelen.org/
“I see that the initial acceleration phase was done by a liquid-fueled rocket engine until about 1960, when it was replaced by a solid-fuel rocket.”
My lab partner in medical school was one of the guys who developed the solid fuel system at Aerojet General. In those days, medical school had a summer vacation from early June to September. He would go out to Aerojet and work the summer. He later went into pathology. In the first year, one of our classes was biochemistry. He flunked the first quiz because the questions didn’t make sense to him. The professor talked to him and he was excused from biochem. He had a PhD in P-chem.
He knew more about carbon bonds than the faculty.
Mike, I think that if it had not been for the development of solid-fueled ballistic missiles…or at least liquid-fueled missiles not requiring fueling immediately before launch…a nuclear war between the US and the Soviet Union would have been much more likely. Windows of a half hour or more to get ready for launch created a situation that was dangerously unstable, strongly tempting both sides toward a preemptive attack.
@David – intriguing thought
Somewhat related this is what I came across today…
Scroll down and you see the original Soviet R1 rocket (1948) was just a rip-off of the German V2
http://io9.com/incredible-soviet-rip-offs-of-western-technologies-973280252
Bill…the US also made heavy use of captured German V-2 technology…we acquired parts of about 100 rockets, assembled them, and fired them at White Sands. The US Army Redstone was largely V-2 derivative:
https://en.wikipedia.org/wiki/PGM-11_Redstone
The Redstone was a V-2 derivative. The first version used the same fuels and the same engine which was uprated about 50% and had better alloys and electronics.
The R1 was a V-2 clone, but the Scud was an improved version of the R1.
The best history of the V-2 and its derivatives can be found at Encyclopedia Astronautica
http://www.astronautix.com/lvs/v2.htm
When I was stationed at Ft Bliss TX, where the Army trains air defense, I learned a lot of the history. The first V2s that we brought back were tested at White Sands NM, about 50-60 miles away. So it was natural that the Army establish their school close by.
At the time the Germans used the Nike Hercules missile too, and I always got a kick out of their barracks – with a V2 in the front.
@David I posted before scrolling up and realized I reiterated what you had said about White Sands.
Those must have been interesting times.
Elon Musk is trying to do for Rockets what Henry Ford did for cars.
Interesting article on him published in Smithsonian’s Air & Space a few years ago – in many cases his is re-inventing critical components, getting the cost down.
http://www.airspacemag.com/space-exploration/Visionary-Launchers-Employees.html
I worked a little bit on the Snark – solid fuel launched ramjet powered bird and ALL of the tracking for it was through the SAGE centers. The “computer” building as I recall was about 250 feet square and 4 stories tall. First floor was power supplies, 2nd and 3rd floors were row on row of equipment racks. Each six inch high bay was four plug-in vacuum tube diodes. EMP was not going to bother them in any way shape or form. “Programming” (if you can call it that) consisted of wiring those flip-flops, and gates, or gates and nor gates according to the Boolean Algebra statement that named the process or computation you were making. Those flip-flops were the “memory”. The Snark was phased out after the Cuban situation and I ended up working Ground Radio for the USAF. My understanding, though, is that SAGE in some form is still functional… I feel quite certain that “Space Command” is still tracking every bit of junk still orbiting around this 4th rock from the sun. Oh, and the “DEW” Line (Distant Early Warning) stretched across the northern reaches and joined up with “BMEWS” (Ballistic Missile Early Warning System) with sites like the Over the Horizon Radar site in places like Thule AB, Greenland.