asp.net - Will a REST interface slow down my search engine? -

To get a search website quickly created, I plan to divide work between two teams: Creating a search engine and creating a web UI (mobile / desktop) is my plan to build search engines as a set of services that rely based on NEAT 3.5. The UI can be created using some other technology.

Question: Is there a possibility of a display interface? What is REST to become an obstacle in this scenario?

It was not clear from your post that you were calling the customer directly from your HTML UI, Or were you calling server-to-server rest on back end? Therefore I will cover both cases below.

If you are being charged between your client UI and your server, then using reverse or other HTM remote viewing approach is relatively less - it takes time to execute the search at the end of the back And then send the results back to the client and reduce the effect of the rest of the call. If you want to improve perf, focus on client-side networking tricks (such as HTTP compression, proper caching header, etc.) and optimize your search engine.

If your architecture is one level of the server (hosting your web UI) then call a different level phone (your search engine), then call again between those parts You should not add more for your overall delay. The reason for this is that the search is going on (as above) and sending the result back to the client usually takes at least a few hundred milliseconds, and the overhead of the back-end cast call (if done properly) usually At 50 mms or less

That said, the customer of server-to-server HTTP calls is easy to mess in the end. For example, many HTTP client libraries (including NATs) will limit the number of concurrent client connections by default, which makes sense if you are building a real client app but will kill your scalability if " Client "which is actually one of the servers provides concurrent service to hundreds of users. Other potential problems include authentication problems, proxy problems, DNS etc. Be careful so that you can carefully prepare and configure your REST client code, and make sure to load-test with a few hundred concurrent users!