A Box of Memories - #Book

New to my blog? You may want to think about subscribing to be notified of new posts. Thanks for visiting!

This is an extract from the book I am writing called 'Living in the Cloud'. It is presented as a draft - comments and criticism are welcome.

A Box of Memories

Although a computer can 'think' about millions of ideas each second, it is important to understand that when you break it right down, it can only think about one thing at a time. Many women may conclude that this is because a man designed it.

This fundamental limitation causes problems even and especially with today's computers, because even though the brain of the computer, the Central Processing Unit (CPU) can operate at blistering speeds, other parts of the computer can't keep up, causing bottlenecks throughout the system.

The Colosseum, the famous ancient stadium in Rome, could seat 50,000 people. It was designed in such a way that the entire ampitheatre could be filled or emptied within minutes, thanks to a clever system of pathways called vomitoria, from the Latin vom meaning rapid discharge. If the Colosseum is the gold standard of architecture for practical and efficient use, the modern computer using the von Neumann architecture is the equivalent of a suburban shopping centre carpark the week before Christmas.

The CPU follows instructions in a four-stage cycle. First, it fetches the next string of numbers in the queue. It then takes a look at the numbers to see what they mean and decide what to do with them. Once it has decided what to do, it carries out whatever needs to be done and then finally sends the results back out to be stored. Then it fetches the next bit of information and the cycle begins again. This is called an instruction cycle. The four steps are usually referred to as fetch, decode, execute and store.

The number of instruction cycles the CPU can process each second is referred to as its clock speed. A CPU with a clock speed of 2GHz indicates that it can process two billion instruction cycles per second. Yes, two billion, that's a thousand million, times two. Clock speed used to be considered the be all and end all of performance, and was certainly marketed as such. Many people shopping for computers got used to the idea that the higher the processor's clock speed was, the faster the computer would be. This is extremely misleading.

In practical terms, a computer is only as fast as its slowest component. When we think 'fast', what we usually mean is 'responsive'. This is also misleading. In reality, a computer is a complicated network of different components with different capabilities and purposes all operating under the control of another complicated system of language with its own idiosyncrasies. But fear not, by the end of this chapter you will have a good understanding of the physical system, called hardware, and the operating system which controls it, called software.

As mentioned earlier, the CPU containing the microprocessor is the brain of the computer. It is where all the calculation, logic and decision-making happens and has the fastest operating capability of the whole computer. The CPU both takes its instructions from and delivers the results to a store of information called the computer's memory. It may help to think of the memory as a turnstile with information constantly passing in and out. The memory also has a clock speed which is typically much lower than the CPU. This is the primary cause of the information bottleneck, since all the information going to and from the CPU must pass through the memory. As you might imagine, there has always been much focus on experimenting with different types of memory technology to try to increase efficiency and clear the information bottleneck.

Memory comes in many different forms. When considering the memory that connects directly to the CPU, it is essential for efficiency that the CPU be able to write to and retrieve information from any part of the memory at any time, at random as it were. Thus it is called Random Access Memory, or RAM. However, the RAM used in most computers stores its information electrically, which means when there is no electricity there is no way to hold the information. When this happens it is called volatile memory.

A computer that loses all its information as soon as it's switched off or there's a power cut isn't much use, so a secondary non-volatile memory must be installed to store the data when it is not in active use by the main memory. There have been many different types of this non-volatile memory storage over the short history of computers, notably the floppy disk, hard disk, CD, DVD and Flash memory card.

Both the floppy disk and hard disk work on the principle of magnetism. In simple terms, a circular disc coated in magnetic film sits on a spindle and spins very fast. A needle sits very close to the spinning disc but not quite touching it. To write information to the disk, the needle changes the magnetism in the disc as it spins by, and reads the magnetism back when it needs to retrieve information.

Looking at today's hard disks, magnetic disk technology has developed to allow very large amounts of data to be stored, but has significant drawbacks. Firstly, because it have moving parts, it is fragile and susceptible to dirt or damage. Most hard disks spin their magnetic platters 7,200 times a minute, so if the needle ends up touching the surface of the disk for any reason it can destroy large amounts of data very quickly. Many people who have experienced hard disk failure and loss of important data know how frustrating this can be. Secondly, because the platters are magnetic, they are sensitive to interference from other magnetic devices. Sticking a floppy disk to a fridge with a magnet is the quickest way to remove all hope of ever retrieving what was on it.

Perhaps the most significant drawback of magnetic disk technology is that it further slows down the information flow. Just as the computer's RAM isn't as fast as its CPU, the hard disk isn't as fast as the RAM. Whenever the computer needs to swap data in and out of permanent storage into active use, it takes time. So what's the alternative? Not CDs or DVDs, which are much slower to read data from than hard disks. It also takes far longer to record information onto a CD or DVD than a hard disk and recording can only be done in huge chunks and not one byte at a time, thus making the 'random access' idea impossible.

The most promising alternative appears to be Flash memory, a random access non-volatile technology much faster, quieter and more rugged than a hard disk because it has no moving parts. Flash memory has been around for a while, developed at Toshiba in 1980. It got its name because at the time you had to clear the whole memory and start from scratch if you wanted to change any information stored on it, a process that reminded its creator of a camera flash. That is no longer the case, and varieties of Flash memory exist in digital cameras, mobile phones, portable music players and gaming devices. It is widely expected that Flash memory, also called solid state memory will eventually replace the hard disk as the primary storage method in computers as prices fall and technology improves.

Nevertheless, despite the varying ways in which they operate, all types of memory perform essentially the same function - to act as a store from which the CPU can retrieve and deliver information as quickly and efficiently as possible. But what sort of information? Where does it come from, and where does it go?