29

Last year at work we started migrating our backend from PHP on a dedicated server to Node.js on AWS lamda functions

We went from 10 second calls to 70ms calls...

At this point our frontend is not even ready for this kind of speed 😅

Comments
  • 0
    Lol pretty much exactly what we're doing rn

    Only difference, our frontends can handle the speed
  • 0
    It's probably a darn sight cheaper, too. Quick running lambda calls are as good as free.

    If only Java could work in the same way without the cold-start times...
  • 4
    Nah, I'm a bare-metal guy. I want full control over what happens and no, I'm not manually configuring them, I'm using Ansible.
  • 0
    Running in to race conditions due to ultra fast page loads after backend optimization? I wonder if that's a first here 😂
  • 1
    PHP is perfectly capable of showing similar times.
  • 0
    @Quirinus Go ahead, run db queries or file operations in parallel in PHP, and don't forget the time each developer will spend setting up their environment before they get all extensions for async operations up & running...
  • 0
    Until you run into cloud formations 200 resource limit. Then you have to start breaking your api up into more and more services. Your backend must be reasonably small. Lambdas a no go for full apis unless truly TRULY micro services. Just my opinion but made that detrimental mistake before.
  • 1
    Php just has to go, it's facts
  • 1
    @dUcKtYpEd that's why we went for the microservices approach :)
  • 0
  • 0
    @shoop phps still one of the most prominent languages you’ll run across. I love node but saying PHP‘s got to go is like saying cobalts got to go along with the infrastructure that 650 billion lines of cobolt code support. Phps prob quadruple that number.
  • 1
    @dUcKtYpEd Source? Best I could find was 200 billion lines in 1997 (when an estimated 5 billion lines were added annually, so 320 billion lines at best by now), with a significant portion of COBOL code being replaced following Y2K (in 2006, only 60% businesses still reported to use COBOL, compared to 80% in 1997, and further 36% had plans to migrate from it in near future or as soon as maintenance costs outgrow migration costs).

    COBOL has been used for crucial business software since 1960, PHP appeared in 1995 aimed at small web pages and home users and it never played a strategic role in core business systems, so it's much easier and faster to replace. Though it's used by a great amount of websites, w3 reports 3/4 of PHP websites run one of approx. 100 most popular PHP CMSs which brings the amount of PHP code down significantly, and while COBOL code is either maintained and in use, or it doesn't run at all, PHP code just remains in limbo on some .com era server in someone's basement.
  • 0
    @hitko you really went down that rabbit hold. 320 billion sounds accurate. I got the number off a red hat podcast months ago so I’m ball-parking from my memory. The reason I say it’s prob triple that (which again is an exaggerated notion to show emphasis) is because I think the amount of website on the webs reached 1.7 billion x .20 (20% of those sites are Wordpress/magneto/drupal/joomla. Prob more then that as Wordpress alone claims 20%) = 340 million. Plus php projects on symfony, laravel, codeigniter and raw php projects. How many lines do each of these projects have of php on average and multiply that by that 340 million. I’m personally all about php going away thoughh. Higher pay for me just like the cobol devs making mad stacks
Add Comment