2.7 KiB
| title | aliases | tags | |
|---|---|---|---|
| software-architectures |
|
# of pieces?
- one integrated piece
- a few distinct components
- many distict components
does it run:
- on one machine
- spread across multiple machines
- in the cloud
these consdier ations dtermine the structure of the software we create
these have changed over time. from monolithic to distributed 
Monolithic architecture
system includes all relevant code in one executable
- e.g., UI, business logic, data management
- self-contained
- no significant separation of reponsibilities
easy to deploy, awkward to upgrade.
- widely spaced "big bang releases" usually only seen in legacy systems now
Separation of reponsibilities
partition systems into logically separated parts. e.g., UI, business rules, data access, utilities
easier to upgrade
each part can be deployed ina different location
each part is more cohesive
- related code all in one place
- unrelated code cleanly separated
- well definec interactions between parts
- easier to design, devleop and maintain
can have "fake data" etc for development
Generic (old) info sys architecture
this is limiting and old. based more of hardware than software.
modern system example
Microservices
everything is a service
- APIs that can be called over the internet
- e.g. product service, customer service
small cohesive units of code
- performs a single function
- each with their "own" data
- not necessarily its own database
- as autonomous as possible
- independent and effectively disposable
- very scalable
- auto matic load management
- can run many redundant instances concurrently
A system is a collection of services
most common implementation architecture for modern distributed information systems and web apps
serverless computing
tradtional web app deployment usually still involves one or more identifiable servers
- includes microservice baed aps
- server may be virtualised
- generally require per sserver configuration
serverless eliminates the "server" concept
- just deploy code into the cloud
- no server config needed
- a form of utility computing
- e.g., google app engine
Function as a service
logical extreme of serverless computing
deploy individual functions into the cloud
- no real concept of a "app"
- fits well with microservice acrchitecture
- very lightwright - fast and scalable
- application logic only ⇒ dont store any data
componentised highly distributed software
ideal for on-demand computing
e.g., AWS lambda, google cloud functions, microsoft azure functions etc