Brandon Miller

While these tools are a great start, you are going to need low latency, and extremely high bandwidth. I couldn't see this happening without you having pretty much your own datacenter at your disposal. While the post below gives some great options, it's extremely apache-biased. Not saying apache is bad, but there are other options to explore. My recommendation is a mix of services from AWS, azure.. Whatever cloud service you choose to go with. You'll definitely need a streaming service.. Either self-built or cloud provided, and a restful API for things that don't quite require the power of a stream. Then, as you scale, you'll start to realize that even the best stacks in the world have bottlenecks, and you'll want to employ use of some type of messaging queue.

With the advent of Gitlab actions/workflows, it's hard to not choose Github anymore. I say that with all love for Gitlab, as it's been my personal tool of choice for a long time because of it's inbuilt CI/CD solutions. Github is just all around more adopted by the community so you'll always find more support; and if you go with enterprise you will get 50k build minutes a month as well as a ton of extra tools that will definitely help a startup out from the get-go. That being said, it's priced at $21 per user, per month so if you cannot afford that, I say go with Github.

Depends on how much time and money the company is willing to invest... But I'd recommend at least getting a start on lifting and shifting to the cloud if you have the knowledge/ability. Lots of performance gains can be made as well as cost saving. Not to mention learning how to operate on the cloud is a great thing to put on any resume these days.

For just a time tracker app? I'd recommend going with a cloud-based approach. A couple serverless functions in whatever language you choose, and the front end can be a static website hosted inside a storage service (blob for Azure, bucket for AWS, etc). This will ultimately probably save you a little time, and them a little money on hosting.

If you want to create a reliable video streaming service, you'll probably want to go with a UDP approach. TCP will throw an exception as soon as a packet drops. That being said..... Node.js is ultimately a bad choice. Way too high-level. I've found, when working with UDP, it's much more performant on lower level languages like C, etc. As far as my recommendation, if you want to go with something 'new' and fun, check out GoLang. It's low level, and developed to handle high performance at scale.

I don't know anything really about CodeIgniter, but I do know that microservices are a great choice for e-commerce, as they tend to have a lot of different moving (but not necessarily connected) parts. That being said, and if you have a choice, I'd recommend Go personally. But node isn't the worst option if Go isn't something you're comfortable with.

If you know typescript and want an entire framework to work with, go wth Angular. AngularJS, to my knowledge was deprecated. React is great, popular, and you'll find a ton of support.

Openstack is made by Canonical, so if you're used to Ubuntu you'll feel right at home in some ways, but it's not so much an OS as it is a cloud management platform. You could use Ubuntu and OS together. As far as choosing an OS, that's really a matter of what you need as a person/company. I will say, I love ubuntu and always use the latest non-LTS versions just to check them out. They've recently took a weird turn away from Xorg and are now on Wayland.... If you need to screen share as a company, be mindful of this because as of right now, you may not be able to when using Wayland.

While it's hard to recommend any framework/library, I'd recommend you start with something that is relatively popular and has a little more maturity. I recommend react because it is arguably the most popular out of the three, so you'll easily find support, and most importantly, a job with this. Vue is a good second option, and also great to learn. To my knowledge, it was actually created by some of the original devs of React. Not sure if that's actually true or not. On to Svelte. This one is actually really great, and I love the approach they took with doing all of the "dirty work" at compile-time. The problem is that it's relatively new, not as mature, and while you're never guaranteed to find a job with any language/framework, your chances are considerably less.

All of this being said, while I do recommend what to start with, just to get yourself into the industry. My personal recommendation for your future career, and just for fun, is to learn them all.