Computer Systems

by Jack Barkstrom - Attorney, CPA

The Internet of Things

The Internet may have started out as a means of communication and information services. It has expanded to incorporate devices, which can check your refrigerator and remind you to re-order eggs or can monitor your heart when you are exercising, or mini-computers, and attachments, or artificial devices (things) and even further. What has been called the Internet of Things (IoT) has now become the Internet of Everything (IoE), which could also include (IoN) (Internet of Nightmares).

ENIAC - The beginnings of the Computer Age

In 1946, the first large-scale electronic computer was introduced. It became known as ENIAC, the Electrical Numerical Integrator and Calculator. ENIAC used approximately 18,000 vacuum tubes and was large, occupying 1,800 square feet (167 square meters) of floor space, and weighed almost 30 tons. It used punched cards for input and output. When it was tested, there was so much uncertainty about the calculations, that they essentially took a vote. Before the computer chip came along, calculations were done using vacuum tubes. If two came up with the same result, they figured the answer was correct. With all the advances in technology things would seem to have changed. Vacuum tubes have been replaced by transisters and reduced in size as computer chips. Yet, the latest innovation in programming and software, Bitcoin and blockchain technology, relies on a vote among participants to verify which chain is correct. As the saying goes, the more things change the more they stay the same.

The first silicon transistor was invented in 1947 at Bell Labs. It was made of two gold electrical contacts and a germanium crystal. Over time the size was reduced and a switch was made from germanium to silicon crystals, which were grown in a lab. The production process involves coating a thin silicon wafer (a slice of the silicon crystal) with a chemical called photoresist. Thin beams of light then cut or etch channels in the photoresist coating which allow electrons to move through the channels. Integrated circuits, which were developed in the 1960s, were the next step, were chips that contained large numbers of transistors. The microprocessor was developed in 1971. It was a more complex integrated circuit, which contained a central processing unit (CPU).

The development of a personal computer market began around 1975. In January 1975 a company called MITS advertised an Altair 8800 kit with an appearance on the cover of Popular Electronics. The kit sold for $395 and required hand assembly, which meant the purchaser had to use a soldering iron to finish the circuit boards. IBM introduced its own version of the personal computer that same year. This was followed in 1976 by the Apple I computer. It was not exactly a commercial introduction, since only 200 of these computers were made and they consisted of a circuit board screwed to a piece of plywood. Its first commercially successful computer was the Apple II, introduced in 1977.

The IBM personal computer introduced in 1975, was not exactly a commercial success, since it sold for $8,975. However, IBM would introduce the IBM Personal Computer (Model 5150) in September 1981. It is considered to mark the beginning of the personal computer and would set the basic standard for personal computers.

The Universal Serial Bus (USB) and minitiarization - An alternative to the mainframe

The personal computer would not have been possible without the development of an alternative to the mainframe. There were two main problems: First, they needed to reduce the size and weight, from huge rooms and tons of weight, to a few pounds and less than a closet, and Second, they needed to eliminate some of the wires. The IBM mainframes, when first created, required a wiring system with direct connections for input and output. We have become familiar with the USB (universal serial bus) cable. Where did that name come from? Originally the "bus" was a single component which was able to manage multiple inputs and outputs. The bus name was derived from "busbar" or "bus bar" which was a metal bar used in electrical substations and industrial cabinets to distribute high voltage electrical currents to multiple devices. For households or buildings, the bus bar was a thick, conductive strip of copper or aluminum found in the breaker panel. It was used to distribute high-current power from incoming service lines to individual branch circuit breakers. Its purpose was to prevent overloading if too much current was supplied directly to household or office lines. Overload would either trip the circuit breaker or, if it did reach the connected lines, short out the system. Both IBM mainframes and early personal computers relied on copper cables for Input/Output (I/O), although mainframe cables were heavier copper, known as "Bus and Tag" cables, which were used to communicate with peripheral devices. Early PCs used parallel ports for printers and external drives.

Up until the 1960s, mainframe systems, which did not have the advantage of microchips, used discrete, individual transistors, which were wired together on logic cards. In the early 1990s IBM developed a fiber optic technology called ESCON (Enterprise Systems Connection). In 1998, it introduced FICON (Fiber CONnection), which relied on fiber optic technology, but was much faster. While FICON remained exclusively a mainframe technnology, its counterpart for personal workstations was SCSI (Small Computer System Interface) cables. As silicon chip technology advanced, personal computers began incorporating this into their latest versions. Intel had developed the first microprocessor (transistor chip), the Intel 4004, in 1971, which fit 2,200 transistors onto a single chip. The Altair 8800, which was introduced in 1975, and sold as a kit, used the Intel 8080 microprocessor.

The single chip, general purpose logic device - The First Microprocessor

The Intel company was founded on July 18, 1968 with the goal of making semiconductor memory both practical and affordable. Silicon chip-based memory did exist, but it was at least 100 times more expensive than the magnetic core memory commonly used at the time. Semiconductor memory (silicon-based) cost about a dollar a bit whereas magnetic core memory was about a penny a bit. By 1970 it had produced a 1Kb memory chip, which for the time, was an incredible improvement, although it only stored 128 bytes, far short of the megabyte (MB) 1,048,576 bytes storage capacity seen today.

Because of its success, it was hired by a Japanese manufacturer Business Computer Corporation (Busicom) to design a set of 12 custom chips for a family of high-performance programmable calculators. It was believed that the logic chips would have to be custom-designed, the then-common design method. Ted Hoff, an Intel engineer, brelieved it might be possible to design a single-chip, general-purpose logic device, which could be used in designs other then calculators. The four-chip set included read-only memeory (ROM), random access memory (RAM), input/output (I/O) and a 4004 processor.

The Silicon Chip Process- Growing the Perfect Crystal

The success of the chip design was only partially due to the creativity of the engineers. Just as important were improvements in the process for manufacturing chips, which involved growing pure silicon crystals. Silicon is the second most common element on earth, after oxygen, found commonly in beach sand. Silicon chips require a very pure form of silicon, so it must be processed, normally starting with raw materials - primarily mined quartz rock, not beach sand, which has more value in the beach and resort market. The process for growing pure silicon crystals, is called the Czochalksi method, named after its inventor. To begin, the quartz rock is heated and converted to metallurgical-grade silicon using electric arc furnaces. The silicon is converted to a liquid, distilled, and then redeposited as semiconductor-grade rods, at a 99.999999% purity level. The next step is the electric crystal pulling oven. The semiconductor-grade rods are broken into chunks, packed into quartz crucibles, and then loaded into electric crystal pulling ovens. The chunks are re-melted at more than 2,500 degrees Fahrenheit.

When silicon crystals "grow" or form, they normally take the shape or structure of the seed crystal. Once the silicon in the crystal pulling oven has reached a molten stage, a small seed crystal is inserted into the liquid and slowly rotated. Some of the liquid will stick to the seed rod, resulting an ingot of crystal, either 200 mm (8 inches) or 300mm (12 inches) in diameter, and more than 5 feet long, weighing hundreds of pounds. The ingot is finished by grinding into a perfect cylinder. Finally, a high-precision saw will slice the ingot into more than a thousand circular wafers, each less than a millimeter thick. To create a chip, a process called photolithography is used to coat the wafer with different layers of various materials, while etching the surface following the design pattern.

The quality of the final silicon chip depends on many variables, including the ingredients added to the seed crystal, which may also be affected by small differences in temperature or the speed of rotation. The goal is to create a nearly perfect crystal and the competition is fierce worldwide. The competition is not limited to the silicon crystal process, but extends to the development of new manufacturing equipment and processes. It also includes development of new coating and etching techniques and the trend toward computerized and automated proceesses.

ARPANET

In 1958 President Eisenhower created the Advanced Research Projects Agency (ARPA), intended to advance the technological development of the military. In 1969, it began research on a project intended to ensure a working network of computers which would continue operations if one of the connections was knocked out, say through a nuclear strike. The original system was called ARPANET, and had only four "nodes" or computer connections, in different cities or university computer facilities: UCLA, the Stanford Research Institute (SRI), the University of Utah in Salt Lake City, and UCSB (UC Santa Barbara).

In 1979, the National Science Foundation (NSF) created a network to connect the computer science departments at universities, called CSNET, using ARPANET technology. In the mid-1980s, the NSF created NSFNET giving other academic disciplines access to supercomputer centers and, in addition, connecting smaller networks together. By the late 1980s, NSFNET served as a basic Internet backbone, allowing high-speed connection points between networks. In 1995, the NSF decommissioned the service and it was privatized, eventually evolving into what the Internet has become today.

Operational Reliability with Sometimes Unexpected Results

Because computers and programs were designed to follow instructions over and over, they were reliable enough to deliver consistent results. This did prevent innovation. Developers were able to create new programs and new uses. There were specific languages, such as Microsoft BASIC (Beginners All-purpose Symbolic Instruction Code), or COBOL (Common Business-Oriented Language). There were programs used to control PC operating systems, such as UNIX or LINUX. There were general categories of programs, such as machine language or assembly language. There were different types of program packages, such as spreadsheet programs, like Excel, which could produce reports dealing with numbers, or word processing programs, which could produce written reports.

There were specialized types of devices and programs, such as the graphical user interface (GUI) which allowed a user to use a mouse to make selections on a computer. There was the development of Hypertext Markup Language (HTML) in 1990, which allowed information to be published on the Internet in an understanble form. Its development is seen as the start of the World Wide Web.

Because of the similarity in design and operation of computer systems and communications, there were vulnerabilities which could be exploited. In 1998, a programmer released his Morris Worm which, while considered nondestructive, nevertheless created problems for about 6,000 of the 60,000 hosts linked to the network. For a time, computer attacks seemed more of a nuissance, but as the Intenet expanded, it became a more serious problem. The attacks seemed to be more malicious, sometimes intionally doing as much damage as possible. Some hackers engaged in spying to obtain information. Others used it to obtain money.

New Challenges and Opportunities

Artificial intelligence (AI) and the development of automated systems and even drone warfare are seemingly new problems, and never-before encountered outcomes. One of the mentioned problems is the potential job loss from artificial intelligence. Yet, job elimination is a problem which goes back to the start of the industrial revolution when factories could mass produce textiles faster than artisan workers, which eliminated some jobs. Andrew Carnegie was constantly trying to automate and streamline the steel-making process, with the goal of reducing the number of workers required. Even computer software packages, such as Excel, have eliminated the need for manual entry and reduced the need for clerks. Over time however, new job opportunities opened up. The new systems were not as self-contained or perfect as claimed, requiring new skills with workers capable of mastering their requirements. New job opportunities opened up as a result.

The rise of data centers to supply the demands of Generative AI training, has threatened to all but take over the energy market. Crypto-mining has also created a demand for, and even theft of, computer resources and energy. These problems are compounded by the general problem of global warming. It is not clear whether a solution has been found for any single problem. New variables have been introduced. Political and economic forces have begun to come into play, with an undetermined impact. Advances in technology have also begun to have an impact. As the saying goes, only time will tell.