Microservices

 Microservices 

Microservices are a solution, to complex online services, normal services used to have 3 principal parts, the html server, a server dedicated to answer all HTML and javascript requests and a database. This monolithic approach is hard to maintain and fragile, if something happens with a functionality in this monolithic services is probable that parts in this service will fail and the load in the service is hard to balance, many people trying to access to only one part of the service could deny that service to other people.


Micro services consist in dividing everything into different components, it allows to have a more explicit component interface. Microservices are meant to b as decoupled and cohesive as possible, basically consisting in smart endpoints and dumb pipes this takes advantage of http-request-response, microservices also give the option to use based on multiple technologies, not limiting your team to exclusively use C++ or node.js for example. The data is not centralized in a single server and database, they let all services to have their own database, cache their self etc.

An interesting thing about micro services is that the division between all modules allows them to fail without shutting down the entire service. If all services are separated it helps to detect failures easily, monitoring in real time each service If a module is not performing well or is failing constantly is easy to replace that module for a new one without shutting down the entire service, in a monolithic approach you need to replace the module and possibly rewrite a lot of code, shut down the entire service and starting the new one whith the changes made. 

Monolithic services are being replaced fast,  but microservices are new and failures and optimizations will appear in a future, the limits for compartmentalization for example and how to develop hybrid systems.



Comentarios

Publicar un comentario

Entradas más populares de este blog

The Hundred-Year Language

 We were supposed to drive flying cars 5 years ago, right?,   The legacy of some old programming languages is something to be analyzed, because these new languages can be better understood if we, as programmers understand how and why these languages came to be and how can they be improved and what characteristics or tools are lacking. A hard decision for other people is to decide the language to learn or maybe what language use for a project that will be need to be maintained in the future. The operators in a language are one of the most important things about a language, these operators are are the axioms of a language,other parts of the language can be replaced or improved relatively easy, but modifying the axioms of a language can drastically change everything that make this language unique and useful, but sometimes this happens.  The languages change, they adopt different things and merge with others, adopting things from others and maybe, this new construction or creature will be
Leer más

Mother of Compilers

 Early computers were more about the engineering process than logic, and given the circumstances of the era, computers, mechanical computers and not human computers were considered something that only male could dominate and understand, but software was something where females could demonstrate their abilities. Ada Countess of Lovelace was the first women dedicated to programming.  She worked on COBOL, one of the oldest but still relevant programming languages and helped computing to be everything it is now, she was in computers when computers were not even a thing, programming was not a thing, programs were something, but the process to input code into a machine was hard and expensive to do.   She was curious when she was young, studying harder than most of other students in high school, and doing much more than was expected for a women.  When the US entered to the second world war, she was assigned to the US Naval Reserve. They knew how smart and valuable to the team she could become
Leer más

Episode 61: Internals of GCC

 As a programmer, you need to make sure that your code will run as best as possible (with a good performance and good compatibility) with the CPU  your program is intended to run, this means that a good compiler for your code is needed, that's where GCC can be used, using GCC you can generate the front end and the back end with little modifications to run in any platform. A compiler will pass multiple times the code to implement everything needed, from the multiple front end implementation that may not be available in C for example, that natively does not support polymorphism, but can be implemented with the correct extensions, or this multiple passes can be used to optimize depending on the platform, operating system or CPU architecture (x86, arm, RISC V etc). Also, the language will delete "garbage code" in the lowest level of your code, will look for small machine code instructions and replace them with faster, better and more secure instructions, increasing the effici
Leer más