CSSE3002 Lecture Notes - Lecture 5: Microservices, Amazon Dynamodb, Graph Database
28 Jun 2018
School
Department
Course
Professor
Software Through the Ages: Architecture Matters
Guest Lecture: Darren Rogan
Talk about: Joys and fun of architecture, reasons behind architecture
Work with business owner, building system
People think: Front end (doesn’t think of backend) doesn’t know what is complex and simple
Twitter: simple, but huge complexity
In the I.T. field, problems always occur, doesn’t know what’s wrong & when it’s going to be
fixed
The value is not what we build, it is what it does for people (we are a service industry)
Will always be asked: What is the ROI? Most projects and startups are based on business
Not about building the best thing possible; it’s about building the best thing within the budget.
And even if you’ve built the best thing and there’s no traffic, it’s a waste too because the
value isn’t there. Business value is more important that how complex your project is,
because u need to get paid.
People think they know what is the final state, but software typically does not look like what it
is supposed to be at the start. The longer the project is, the further the deviation from the
initial design or concept of the software. Reason: pivoting, new requirements etc
Monolith apps: giant singular code base with every functions and feature. Why? It was easy.
Can build it on one single machine, straightforward.
Problem of monolith: can only scale it by giving more resources, afraid to make any changes
on the code, afraid that it will destroy other parts.
Multithread programming: good, but actually not that good. Why? Deadlocks. More threads
more problems.
As the project gets bigger, more investment in time, money and effort but less profit.
Most companies are not tech comp but business company.
Those technology companies build their own technology.
Generally underfund operations. Operations doesn’t necessary have a definite business
case.
Double revenue, stop paying for operations, push to next release
People: Throw everything that’s difficult into operation problem, finish project and run away.
Problem still exists.
First project: make most money, second project gives it mostly for free, next project has to
buy a bigger server (hence use cloud)
If need a little bit storage: no need to buy hardware as server, if not there will be unused
resources
Monolith problem: problem is cost of change and risk of change
find more resources at oneclass.com
find more resources at oneclass.com
Document Summary
Talk about: joys and fun of architecture, reasons behind architecture. People think: front end (doesn"t think of backend) doesn"t know what is complex and simple. In the i. t. field, problems always occur, doesn"t know what"s wrong & when it"s going to be fixed. The value is not what we build, it is what it does for people (we are a service industry) Most projects and startups are based on business. Not about building the best thing possible; it"s about building the best thing within the budget. And even if you"ve built the best thing and there"s no traffic, it"s a waste too because the value isn"t there. Business value is more important that how complex your project is, because u need to get paid. People think they know what is the final state, but software typically does not look like what it is supposed to be at the start.
