New England Data Services Blog

What Is Cache?

So you’re sitting at your desk to doing some work. You are just about to sign a contract, but there is no pen on your desk. Remembering that the company keeps all the pens 3 floors down in the warehouse, you jump up to go get one. Finally, you arrive back at your desk with a pen, sign the contract and return the pen to the warehouse. Ten minutes later, another contract comes in and you need the pen again. Now you have to go back down 3 flights of stairs, get the pen, and bring it back to your desk. I get tired of people explaining how computers work with all the technical terms, so I will continue to use this office situation throughout this explanation.

 

Essentially, this is how a computer works between the guy working at the desk (the CPU) and the warehouse (RAM) where your necessities are stored. Unfortunately, the design of a computer does not allow the RAM to be right next to the CPU. When the CPU must perform a task, a signal has to travel from the CPU to the RAM, locate the desired action or information stored in a memory location, and bring it back to the CPU where it is processed. Wouldn’t it be nice to have a storage place right on the desk to keep pens?

 

Cache is a small memory system built right on CPU to eliminate the need to travel to the RAM for frequently used processes or information. Cache is a very small size and does not hold large files. A small and frequently used process like adding two numbers is a task that can be stored on the cache. In our office example, something like a pen could be stored on our desk (cache) because of its small size and frequent use.

 

Sometimes your desk just gets too cluttered and you need to clean it off. This occasionally needs to be done with cache memory. Filling up your cache slows your computer down and makes it so the CPU has to travel all the way back to the RAM again. If your desk is filled up, you can’t really find anything and you have to go back to the warehouse to get another pen.

 

You may be asking why the cache doesn’t just store all of the information of the computer so RAM is not needed, right? The reason is that cache is very expensive. In other words, we use the small amount of cache to speed up the larger amount of RAM. In other words it is much cheaper to use a warehouse for storage than buying desks to fit everything in your warehouse.

Compressed Files: How They Work

Have you ever downloaded a file from a fellow co-workers email with a .ZIP extension? When you open the file up, WinZIP takes over and creates a folder with a collection of text files that he has been working on. You also notice that the size of the actual .ZIP file was smaller than the files that came out of it. The laws of physics cannot explain how a large object can fit in a box physically smaller than itself. The trick to doing this is through file compression.

 
Here I will take a quick phrase and compress it. 

“Try not to become a man of success but rather to become a man of value”

 This phrase contains 16 words, 55 letters, and 15 spaces. All together we will say this phrase takes up 70 (characters and spaces together) units of memory. We see some redundant words that we can create a common value for. Using the following key, we can assign some words to our common value. 

 Now our sentence reads:

            “Try not 1 2 3 4 5 success but rather 1 2 3 4 5 value”

 When we assign the common values to certain words, we have come out with fewer characters. There are still a total of 16 words and 15 spaces, but now we only have 37 characters. Together we can save the new phrase in 52 units of memory. If you knew the system we used to compress the original phrase, you could easily translate our compressed sentence into the original phrase. Essentially, this is what your compression software, WinZIP, does when you double click on a .ZIP file.

 

Searching for even more patterns we can see the string “1 2 3 4 5” appears twice. Assigning those both to the number 6 eliminates having to double save “1 2 3 4 5” into memory locations.

 

            “Try not 6 success but rather 6 value”

 

Now we have an even more compressed sentence with 8 words, 29 characters, and 7 spaces, making 36 units in total. To recap it all, we have taken the original phrase with 70 unites of memory and compressed it down into 36 units of memory. The computer now takes the 36 new memory units and the compression algorithm and zips it all up into a .ZIP file.

Even though I saw patterns in words with this sentence, patterns of characters can also be put together and assigned to a common value. For example “cc” or “at” could appear in the text after or before this phrase within the file. There is countless number of ways that your software can take to find redundancy in a text file. This works excellent with text files containing large character strings. The larger the file size, the greater chance of redundancy. File types, like video and audio files cannot be compressed because there is little redundancy within the file type.

 

Our above example shows the type of compression called lossless compression. With this type of compression, the file you compressed will be the same size as the file you get when you decompress it. It is like physically taking a big object of X volume, breaking it down into pieces and squeezing it all into a box of smaller volume with the instructions on how to piece it together. When you open the box you read the instructions and rebuild the object. The other type of compression, lossy compression, is very different. When breaking down the same object, it is like putting the pieces that fit in the box and throwing out the extra. With fewer pieces than you started with, you cannot rebuild the object that you started with. Applying this example to audio and image files, lossy compression results in quality lost. An image of grass with a lot of color and texture may be compressed into a solid green image.

 
After reading this blog, you will never look at a compressed file the same way. We can now see how useful compression is when sending large files through an email document or file sharing website. I hope you don’t go hacking that $40,000 vase apart when you have to send cross country it in a physically smaller box.