Computer Architecture

• Main components of a computer
• Disk / Storage: Hard Disk Drive, or Solid State Drive (faster) — Stores all of our data
  • Persistent — Even if our computer crashes, the data stored in here should be OK
  • How much data can it store?
  • Bits: Can store either 0 or 1
  • Byte: 8 bits
• RAM (Random Access Memory): Also used for storage (usually alot smaller than Disk, i.e. 4gb ram, or 8gb ram, etc) but it’s alot CHEAPER in terms of RESOURCES
  • The amount of time it takes to write data into RAM is VERY VERY fast compared to writing to a disk!
  • Reading and writing in ram is measured in microseconds, because it’s that fast, vs milliseconds for disk —> At least 10x faster, or sometimes a 100times faster!
  • RAM storage is not persistent (saved) where-as disk storage is!
• How do we read or write data to our ram or disk?
• How do they communicate with each other? Is this even possible (yes)?
  • Introducing… CPU
• CPU (Central Processing Unit): Writing and reading information from RAM and into HDD / SSD
  • What exactly does it do with that data…?
    • CPU will execute code soo.. that code is stored somewhere in RAM
    • Code that we write in english is complied in 0s and 1s..
  • CPU also writes / manipulates / doing computation (Arithmetic, adds/ substracts, etc.. B/C everything is represented as 0s and 1s!)
  • What if we want to go even faster? — Enter CACHE
  • Cache: A part of CPU which is smaller than RAM, but part of CPU
    • This allows us to write EVEN faster (nano seconds!)
    • CPU chooses which part of the ram we are using very frequently and puts it in the cache — so this allows us to read/write/ compute even faster!
    • So next time we have to read it.. It goes first to cache to check, and then just
• Downside of CACHE / RAM: not persistent data! Unless you save to HDD/SSA, it has a possible chance of losing all of that data
• We take all of what we learned (machine) and combine multiples of them to make a distributed cluster

Application Architecture

• Production App — How does it work?

• Server - A computer that can handle requests (able to serve users)

  • Server can also talk to other / different servers. FOr example, if you ask a server to make a payment, it might invoke an API call to the Stripe server which handles online payments

• From a Dev’s perspective, a simple app usually involves the following components: Building and deploying code, server, storage, load balancers, logging, metrics, alerts, etc.

  

• A user would have to send requests to Server, and our API responds in some form of answers/ data

• If we have alot of users keep requesting stuff from our server, we run into computing limitations (RAM, CPU, memory… etc) → Note we could always upgrade the server! (Vertical Scaling)

• We can also do Horizontal scaling (Having multiple servers instead of one — Note in horizontal scaling, the servers are of the same size / capacity)

  • Multiple servers → User’s request should go to which server?
  • We need a load balancer to tell us which server has capacity, so it can send the request to that one

• As a dev, we also have some external service which is used for logging (Stores all requests if they went thru, had an issue, etc)

• Another external service is Metrics / Monitoring → How our application is running (Some metrics might come from logs)

  • CPU / RAM useage overtime …
  • How many requests were timed-out

• We can use the metrics to set-up Alerts → This allows us when something is wrong! For example, if CPU reaches 95% capacity or about to be full.. Our Alerting service can automatically immediately let us know

Design Requirements