Odtnhj

I had a requirement of implementing a basic rate limiting for a simple api, (with max 40- 50 requests per day) in Java
requirement was:
Request from an IP should be limited to N in M seconds (eg: 10 reqeusts in 60 seconds)
Here is what i did:used a filter and the logic was:
Approach 1:
used cache2k for saving the ip address and a pojo with counts and last accessed time

put each ip as as key and pojo as value (pojo has fields such as :count, last accessed.)

once a requests comes, check the ip in cache,

• if no ip retrieved, insert into cache:
  key:ip
  value:pojo with count 1 and last accessed as current date.



• 
  

  if the ip is retrieved, then check last accessed date.

  
  
  

  • if difference between last accessed time and current time is > M seconds that means ,no requests have come from this ip in more than M seconds. create a new Pojo with count 1 and last accessed as current date and insert to cache

  
  
  • 
    

    if difference is < M, then check the count .

    
    
    

    • if the count is < N ,then increase the count in the pojo and re insert the value with last accessed as current time and allow the request.

    
    

    • if count is = N, filter the request- don't allow the request

    
    
  
  

Approach 2:

• Set the cache expire time as M seconds :



• 
  

  Check if the entry is there in cache or not.

  
  
  
  • If not , insert key as ip and value as count(which is 1)
  
  
  • if yes, check the count ,
    
    
    • if count is = N , then don't allow
    
    
    • else, insert the value with updated count and allow the request
    
    
  
  

I don't want to over complicate things by using Guava for a simple application.

Would be great if some one can give suggestions.

Thanks,