Blog‎ > ‎

Working Skills, Interests, Garbage Collection, Azure, AWS, Serverless, HTTP/2 and Random Studying

posted Dec 16, 2017, 11:29 PM by Sami Lehtinen   [ updated Dec 16, 2017, 11:30 PM ]
  • On work market it's important to understand and be able to utilize different software tools, like Photoshop, Excel and Wordpress, Python, SQL databases, etc. Programming, Scripting and Automation is also highly valuable. Fluent English is also very valuable, especially when working with international projects. Search Engine Optimization is also quite nice skill, when and if you're designing web pages as well as social network and marketing skills.
  • Just read my blog if you're interested. What have I done in past. Different kind of tasks, software, projects and general opinions.
    Like: Servers networking programing consulting Web maintenance databases programing software security and physical security access control. Integrations specifications, project management, disaster recovery, backup management. Bug hunting, accounting audit trail stuff. Nav and SAP integrations. Lot's of SQL and transactionality stuff, data integrity. Statistical analysis and processing raw data. Gathering it. Writing incident and issue reports that vendors can't dismiss with their generic bs answers.
  • Julia provides a litmus test for job descriptions. Truly classic. Because job descriptions are so often so ridiculous, that's why my job description is now just "Techie". I do stuff, which seems important.
  • Quantifying the Impact of "Cloudbleed" - A very good post about it in Cloudflare blog. As well as researching the actual effects it caused.
  • Checked out Bose–Einstein condensate. Superfluid, Supersolid.
  • Read: Azure in Plain English and AWS in Plain English. It's nice that everyone got their own naming conventions for everything.
  • Read: Modern garbage collection - A very nice post about GO garbage collection / colletor (GC). Nothing new in the article, to sum it up, everything is a trade-off. "concurrent, parallel, compacting and generational" - Python 3 Garbage Collector interface documentation.
  • Read: Real World HTTP/2 - Yep. HTTP/2 optimization is complex, and requires also server to work in efficient way. It's one part of the stack.
  • Read: Serverless Architectures - kw: FaaS, BaaS, PaaS. Hmm. Kind of funny, but this is all the same, just bit different isolation, and restrictions etc. FaaS reminds me good old CGI so much that it makes me laugh about all this hype. I just deployed while ago an CGI application on one Windows server, which runs very seldomly. Yes, it's an application, which requires no administration. It uses files in the application path, and is loaded only on demand by the server. Ha. Yes, in this case, there's no automatic 'deployment tool', but there trivially could be. Again, it's just RPC, either it's your or service providers code. But it's code which is run on demand and potentially lazily loaded and evicted when not required. All very basic stuff. Or any other old fashioned 'web hosting' which allows you to run arbitrary code on their server(s). They're responsible for the environment, databases, etc. This is just automated and standardized environment. Not very different from fully hosted LAMP stack in logical sense. Also Tomcat was very very similar, when someone else administered the servers and environments, and you only used application with MySQL or so. Automatic horizontal scaling is of course different from that ,but it's the pro of just fully automated instance generation and resource allocation. It's illusion that applications will magically scale to any scale. It's very easy to ruin all this by locking and any shared sequences, counters, etc. Of course that's bad design, but it happens very easily if not being considered during the application design. I'm not against FaaS, but it's just more hype. Would I use FaaS, yeah sure. It's very similar to App Engine PaaS. Which is awesome, except from pricing stand point.