| Academic Computing and Communications Center | ||||||||||
|
|
|
|
|
|
|
|
|
|
|
|
Operating Systems Support Group |
||||||||||||
|
||||||||||||
| A Brief History | ||||||||||||
|
A brief history of the other half of the
UIC Academic Computer Center's services – its computers – and an introduction
to the Operating Systems Support Group that keeps them going.
Formerly known as the Basic Systems Group, the roots of Operating Systems Support Group of the Academic Computer Center reach back to the Center's founding in the late '60s. As a matter of fact, the Computer Center originally consisted of just two divisions, Basic Systems, which took care of the hardware and operating systems, and User Services, which provided consulting. The traditional role of the Operating Systems Support Group has been to ensure the reliability and availability of UIC's centralized, campuswide computing facilities. Our group has a mission akin to that of the plumbers, electricians, and other UIC service staff who maintain utilities on campus: keep things running smoothly. And, as with the plumbers and electricians, our work goes mostly unnoticed. We only attract attention when something has gone wrong and people can't access Computer Center services! Our basic duties boil down to three:
|
||||||||||||
| A Look Back | ||||||||||||
|
A look back at the history of the Computer Center will help explain how
our group evolved from a "few" people serving a tiny base of scientists
into "fewer" people who now support critical services on many servers used
by virtually everyone on campus today.
Return to Contents Getting Started: the '60s through the '70sThe Computer Center began back in the late '60s. Our first "computer" was an IBM model 1620. The 1620 wasn't much more than a very expensive and very slow calculator that used punch cards for data and program storage. It did arithmetic via a table lookup mechanism: to compute 3 times 7, the hardware looked in row 3, column 7, of a table and got a 1, with the carry of 2. Because the table was in RAM -- which was "magnetic core" in those days, planes of small magnetic "donuts" with wires passing through the holes -- you could accidentally alter multiplication results if you were sloppy with your programming!At that time, we also had an IBM 1800, with peripheral devices intended for process control, analog-digital converters, and the like, which ran an operating system called 1800 MPX (Multi Programming eXecutive). In 1968, the IBM 1620 was replaced by an IBM 360 Model 50, which was the world's first popular commercial mainframe computer. It had four disks drives (7 MB each!) and 256 KB of central memory. It ran an early IBM operating system called MVT (Multiprogramming with a Variable number of Tasks; the "tasks" were programs). While punch cards were still the main method of program input, the 360 supported CPS, the Conversational Programming System, which allowed three users (wow!) to write and run interactive BASIC and PL/1 programs on state-of-the-art 130-baud IBM Selectric typewriter terminals, without leaving their seats! The 360/50 was shut down for about 10 days in May, 1970, during local campus student riots that followed the murders at Kent State. Soon after we got an IBM 360/65. It had extra memory -- 1 MB of it! -- in a box larger than a refrigerator. If you looked closely, you could actually see the bits through a big glass window. It cost over $1,000,000 even though we just rented it. CPS was still there, along with TSO (Time Sharing Option), which allowed more interaction, but was a slow kludge on this non-virtual memory machine. Your terminal would type out logon proceeding... every 30 seconds for a very long time until you got its first prompt. Next came an IBM model 370/155, which was soon upgraded to an IBM 370/158. The 370/158 was our first machine with a "virtual storage" memory system. The Multiple Virtual Storage, MVS, operating system allowed programmers to access 16 MB of pretend memory that was mapped into the machine's actual 1 MB of early semiconductor memory. We still run a version of MVS today to support the library's online catalog UICCAT. In the late '70s, punch cards and keypunches were replaced by terminals and SuperWylbur, a CRT-based keypunch replacement for program development and submission. Although not too many people noticed, during the 1970s the UIC Academic Computer Center developed one of the most available and helpful campus computing services in the country. All faculty, staff, and students, regardless of their affiliations, could use real mainframe timesharing systems and not pay a fortune. Universal access to computing facilities was very rare at the time. Return to Contents The '80s to Mid '90s -- the CMS YearsBy the early 1980s, it had become clear that what we needed was truly interactive computing. The Computer Center addressed this need by implementing IBM's new VM/CMS operating system on an IBM 4341 mainframe computer, which had a whopping 12 MB of memory. The 4341 was able to support 200 simultaneous users, who could develop and execute programs without using cards or keypunches. As the '80s progressed, CMS usage and resources grew exponentially and we made several upgrades to our mainframe facilities. We added a second 4341 and ran CMS on one and MVS on the other. In fall of 1986, we moved our machine room from our SEL facility to BGRC, and took the opportunity to replace our IBM mainframes with an IBM 3081. In 1990, we got an IBM 3090 model 120 with one CPU and a vector processing facility. At that time, our 3090 was considered to be a (small) supercomputer! It has since been upgraded to a 3 CPU model 300J, with 256 MB main memory and 512 MB expanded memory for paging. We still have the 3090 today, still running VM/CMS.Around 1982, the ADN's mainframe was connected to the fledgling BITNET network, providing our first email connectivity to the outside world. In 1987, UICVM became a node on the relatively unknown "Internet" network. The UICVM mainframe was soon the center of network communications for many users on campus. As the 1990s dawned, Internet and its associated activities were growing dramatically and the UICVM mainframe was upgraded to meet demand. At peak usage during the mid 1990s, the system was providing service to over 12,000 members of the campus community. At times, 1000 people were logged into the system simultaneously. Return to Contents The 90s and BeyondThrough the late '80s and early '90s, market forces, ubiquitous personal computers, and low priced UNIX-based "network servers" were drastically changing the rules of the game for computing and communication. The cost of operating traditional mainframe computers was growing and their utility was dwindling. The Computer Center, however, had become so focused on providing excellent mainframe services that we failed to notice this great wave of change for a while. But we began to catch up by the mid '90s, when we made major alterations in our direction to better support personal computer users and their ever-growing demand for network connectivity and services.Fortunately for some of the older members of the Operating Systems Support Group, the Computer Center was willing to spend the resources necessary to teach us old dogs new tricks, rather than finding new dogs. In the summer of 1994, we old dogs underwent baptism by fire, resulting in the roll-out of tigger, our first general purpose UNIX server, for use by UIC faculty and staff. The rush to get tigger into production was made all the more urgent when programmers at UIUC/NCSA developed Mosaic, the first viable World Wide Web browser. The Internet, which had been used almost exclusively by scientists for nearly two decades, was about to burst onto the world technology scene with the same impact as the invention of the telephone, electric light bulb, and the automobile. The original tigger was an IBM RS/6000 model 590 with one CPU and 256 MB memory, running AIX, an IBM variant of UNIX. Tigger is so popular that it already has been upgraded three times. The current tigger is an IBM RS/6000 model J40, with six 112 MHz PowerPC 604E CPUs and 512 MB memory; tentative plans for this summer call for upgrading tigger for the fourth time by replacing it with a new machine with three times the computing power, memory, and I/O bandwidth of the current tigger. (As with previous tiggers and icaruses, we'll run an ADN service or two on the old machine. The original tigger, for example, is now home to ness, our Bluestem ID server.) In the spring of 1995, we introduced icarus, a general-purpose SUN SPARC 1000 server running the Solaris flavor of the UNIX operating system. Icarus is dedicated to serving the networking and computing needs of our student population. Icarus, too has been upgraded several times; its current incarnation is a Sun Ultra Enterprise 4000 server with 1.5 GB of memory and five CPUs. Over 25,000 students currently have accounts on icarus; most use it for their personal Web pages and for email access. To keep up the trend of a new UNIX server each year, in the summer of 1996 the Computer Center purchased a "compute server" to provide high-end computing power for campus researchers who had outgrown our and their workstations. Borg also has been upgraded a number of times; it is now an HP/Convex SPP1200 system with 16 processors and 1 GB of memory. Plans are in the works to replace borg this summer with a machine that will have almost four times the computing power of the old machine and 4 GB of central memory. (That's over 16,000 times as much memory as our original IBM 360 mainframe!). We also maintain several other servers that provide @uic.edu email routing, Domain Name Service support, LISTSERV discussion groups, Netnews/Usenet newsgroup services, printing in the ADN microcomputer labs, ADSM workstation backups, and even audio and video streaming services. And, just to remind us that it can often be just as difficult to phase out old technology as it is to implement new, we still run those MVS and VM systems that were started in the 1970s and 1980s! We hope to phase both out early in the next millennium. (See A Time of Opportunity.) Return to Contents
|
||||||||||||
| Thirty Years of Computing | ||||||||||||
Return to Contents |
||||||||||||
| Three Truths | ||||||||||||
As we moved along the UNIX server learning curve, we discovered three truths:
|
||||||||||||
| The Group | ||||||||||||
The Operating Systems Support Group, with nearly 80 years of combined systems administration experience, consists of the following seven people, listed in order of UIC seniority: Ed Zawacki Michael Homa John Schulien Steve Jordan Roger Deschner Jon Salud (young and energetic) Jim O'Leary, systems@uic.edu Special thanks to George Yanos, former associate director of the Computer Center, for providing some details on the early history of computing at UIC. Visit the Operating Systems Support Group on the Web at
http://www.uic.edu/depts/accc/systems/
Illustrations (c) SoftKey International
Inc. and its licensors.
|
||||||||||||
| The ADN Connection, April/May/June 1998 | Previous: Copyright and Fair Use | Next: Guidelines on Email Size |
| 2000-1-19 connect@uic.edu |
|