Why Big Tech Infrastructures Are Usually Comprised of Self-Picture Solutions

Cool. article CTO of the Yandex Platform Engineering team, in which he explains the emergence of custom solutions in large companies. He highlights. 6 reasons:

1. Innovation Large companies often face new problems for which there are no standard solutions. And then they make those decisions themselves.
2. Scope This item is partially embedded in the previous one, since often new problems are due to the scale of the company and its tasks. As a result, default decisions on such a volume do not work very much.
3. inertia When your own tool is already made and sharpened to solve an internal problem, then when a common tool appears, it is often not so well sharpened for specific cases. As a result, moving often requires finishing the common tool itself.
4. Migration This is about the cost of moving users. Often, such migrations from old tools to new ones are expensive and can last for a very long time, which undermines the very meaning of moving - it is cheaper to continue maintaining your own bike.
5. Business continuity Internal tools allow you to ensure independence from vendor solutions and even from changes in the licenses of open source projects. What is important in the current reality
6. Syndrome not invented here (NIH) This syndrome affects the decision to saw your own bicycles and this is an objective reality.)

The author demonstrates these reasons in two examples. Yandex’s own cloud, which appeared before K8s and now continues to be used and developed Monorepository Arkady, which allowed to solve the problem of svn, git, mercurial and still be able to store all the code (2 TiB) repository

Interestingly, I wrote about it in an article.How RnD Appears in Big IT CompaniesGoogle, Amazon and why Tinkoff needs it:)

#Management #RnD #PlatformEngineering #Engineering #Software #SoftwareDevelopment #Leadership