IBM recently announced its sixth annual Master the Mainframe Contest for high school, college and university students across the United States and Canada. The contest is designed to get the younger set away from their iPhones and Droids and get involved with some serious mainframe applications.
According to IBM in order to win competitors have to complete a series of extremely difficult and time-consuming tasks, mastering the inner-workings of mainframe software products, security protocols, multiple programming languages and various utilities. In the past students had three months to explore the system and complete the contest tasks. The 2009 IBM Master the Mainframe Contest drew over 3,000 students from 425 schools.
Statistics about the graying of mainframe experts and a shortage of new talent to fill their shoes are nothing new. Last year, the AFCOM Data Center Institute reported that more than 60% of IT workers with mainframe experience are now at least 50 years old.
The competition comes on as a new generation of IBM mainframes rolls out: The zEnterprise. With that in mind it's worth taking a look at key points in the evolution of the Big Iron:
The beginning: IBM and others have been working on large systems for many years. In IBM's case you can go back to its transistor-based 7090, which was rolled out in 1959. But it was in 1964 that IBM began what many would consider the first true series of mainframes, the IBM 360. IBM at the time said the System 360 "includes in its central processors 19 combinations of graduated speed and memory capacity. Incorporated with these are more than 40 types of peripheral equipment which store information and enter it into and retrieve it from the computer. Built-in communications capability makes System/360 available to remote terminals, regardless of distance. The equipment is supported by programs which enable System/360 to schedule its own activities for non-stop computing that makes most efficient use of system capabilities. The system's machine cycle time -- basic pulse beat of a computer -- ranges from one millionth-of-a-second to only 200 billionths-of-a-second."
The cement: While the System 360 may have ushered in the mainframe age, the System/370 launch in 1970 cemented it. At its height of its popularity IBM said there was sometimes a two-year waiting list for the machine. According to IBM, monthly rental for a typical System/370 Model 155 having 768,000 bytes of main memory is $47,985, with a purchase price of $2,248,550. Monthly rental for a typical Model 165 with 1-million bytes of main memory is $98,715, with a purchase price of $4,674,160. "The applications of the Seventies will involve increased multiprogramming, remote computing, management information and tele-processing networks that make a centralized computer data base available to people in many places," IBM stated.
Getting big: First introduced in 1985, Enterprise System/3090s had over 1-million-bit memory chips and featured what IBM called Thermal Conduction Modules to speed chip-to-chip communication times. Some 3090s ran something called a Vector Facility that was designed to "increase the computing performance capability of the central processor when iterative or repetitive logical or floating-point operations are to be performed on many related data elements," IBM stated. Prices for larger models topped $5 million.
The next big thing: In 1990, the mainframe got a serious refresh with the 18 model, System/390 rollout. The S/390 also brought with it IBM's high-speed fiber optic mainframe channel architecture Enterprise System Connection (ESCON). IBM said the S/390 featured ultra-dense circuits and circuit packaging that allow higher performance, integrated encryption/decryption for sensitive data, extended supercomputing capabilities, and twice the processor memory previously available. Keep in mind too that IBM touted its heritage by noting that programs and applications based on the widely used S/370 architecture will run on the new systems with little or no modification. Basic purchase prices for the water-cooled models range from $2.45 million to $22.8 million, IBM stated. IBM updgraded or added about 60 models to the S/390 line throughout the '90s, with a major refresh hitting in 1999 with the introduction of the S/390 Parallel Enterprise.
Not all about hardware: In 2000 IBM said it would begin supporting Linux on the mainframe, a huge nod to that systems growing influence. By 2009 70% of IBM's top 100 mainframe customers were running Linux, according to Gartner.
All about the zzzzs: Saying it had rebuilt the mainframe from the ground up, IBM in 2000 rolled the eServer zSeries out. IBM said it spent two years and $1 billion to develop the machines. IBM said the key to the new mainframe was its multichip module (MCM) -- the densest, most advanced semiconductor and packaging technology in the world. The 5-by-5-by 0.25 inch module contains 35 chips mounted on 101 layers of ceramic glass connected to 4,226 I/0 pins by 1 kilometer of wire. The main machine ran 2,500 MIPs on 16 processors and if clustered could handle up to 9 billion transactions/day (300 million transactions/day stand-alone), IBM stated. In 2003 IBM introduced the z990 and called it "the world's most sophisticated server."
Moving on..again: Recently IBM introduced a new generation of mainframes that include the ability to manage power and x86 IBM blade systems from the mainframe consol. The zEnterprise can have up to 96 processors and will support up to 114 blades with eight cores. With their characteristic excitement, IBM mainframers call the zEnterprise the most significant change to the mainframe platform in 20 years. IBM says the new mainframes run some 40% faster than its predecessor but uses 90 % less energy to achieve that performance. The key to this new generation of mainframes is its ability to plug other computers into the mainframe and manage hardware and software applications, according to a New York Times article. In theory, they say, that should make it possible to extend mainframe strengths in security and always-on availability to other systems. Certainly the new design is intended to make it possible to blend tasks, parceling out various parts of a complex application to different computer systems, more efficiently than in the past, the article states.
Follow Michael Cooney on Twitter: nwwlayer8