Interview with a Boomer CTO in 2023

oh yes please please [Music] let me ask you have you used AWS you need whole team to set this up now you can't just use Heroku if you want to have scalable projects you're going to want to think about kubernetes Cassandra Kafka spark I have 35.5 year in the industry experience I have done everything huge mobile apps huge local only web platforms data integration data processing data knowledge Big Data integration systems holistic [Music] um holistic Big Data integration Edge computed next-gen hybrid native systems huge deficient data migration distributed monolith platforms Cloud computed cloud services a lot testing for data centers data center testing for load centers everything from a Raspberry Pi to IBM Watson from pacemaker firmware to nft scam I have used all used databases and neo4j last year I did huge data migration project and you know what it costs you know what it costs 180 000 AWS credits that's why we get paid the big dollars of course it was never launched to the real world only corporate internal project let me give you metaphor Facebook hear me out we want to store the data fail safe redundant at the Hybrid native cross performance distributed attached in the cloud redundant you got to think about processes security practices that data integration deployment pipeline processes radies Jenkins postgres elasticsearch continues integration git actions won't get you very far you need whole deployment team devops team you need devops guy available on call 24 7. another guy for the rest and the third guy because you can't trust the second guy listen JavaScript you want to use something like Java you want to use something that's really statically typed file safe like Java not like rust rust is playmobil right now no integration no adoption you need testing in place I'm talking about white box testing Black Box testing Hot Box testing non-functional testing endurance testing spike testing availability testing stress testing test test testing UI does not matter at this stage you got to think about processes you got to have processes in place you want to first get your requirements correct you need requirements manager requirements managing process you need whole team just for requirements acquirements and you need whole team to control that team another team to control that control team you need whole team for deployment pipeline another team for integration pipeline you need testing systems in place v-model code implementation comes at the very end usually done by one person agile agile agile is no good here you want something that's stable you need processes in place I will tell you what you do with agile you look at your code you look at the trash bin you throw in the trash pin microservices and then your code looks like Twitter if you don't plan to sell your company you don't use children's magic like microservices what is this processors verification validation the model business requirements systems requirements high level low level high level design coding unit testing functional testing integration testing scalability testing system testing acceptance testing yeah but it works it's not non-functional requirement you need maintainability a reliability scalability localization user acceptance usability compatibility performance scalability acceptability reliability maintainability portalability quality safety Mamma Mia what is documentation validation you need verification if you don't do that you know what you do you're doing code transification what can I do everybody uses file Safe Systems Amazon S3 Uber S3 Netflix S3 no you can't just use that you know what you're going to do uh you're going to take a S3 Docker kubernetes you need whole team to set this up and we did not even get to talk about databases let's not get ahead of ourselves load testing spike testing UI comes after concept stage I know I'm making all of this sound very simple but you shouldn't be there to fix that you should be doing all of this in concept stage V model in the model you go for water step you go back a step you fight a little bit and so on and so forth until user acceptance testing repeat you will need a working plan remember visual Services startup and then your code looks like Twitter am I being too elaborate what you got to do is take processes processes process processor processes we mode data integration pipeline requirements acquirements processes security practices testing deployment pipeline Dev of steam whole testing team testing processes in place processes processes operational acceptance testing or a team repeat acceptability scalability depth of steam I know I'm making this all sound very simple but you shouldn't be there to fix that no problem no problem yes goodbye

Channel: Programmers are also human

Views: 301,409

Keywords: AWS, SDLC, cloud services, programmer humor, security practices, startup humor, tech humor

Id: nqa_Uyz1pBE

Length: 5min 59sec (359 seconds)

Published: Sun Jan 29 2023