d********w
发帖数: 363 | 1 Implement rate limiting for an API so that we rate limit if the same
developer makes > 10K req/sec. What would the logic be? What would the
structure be? Where would you store this structure?
一家SF公司的面试题 | d********w
发帖数: 363 | 2
我想这个设计
map>
if (map.containsKay(developId))
last_time, last_count = map.get(developId)
if last_time != cur_time
map.put(developId, new pair)
else {
last_count++;
if (last_count > 10K)
return false;
map.put(developId, new pair)
}
后来又问如果是多台机器（后端做load balancing）如何限制总流量，大家觉得如何考
虑？
【在 d********w 的大作中提到】
: Implement rate limiting for an API so that we rate limit if the same
: developer makes > 10K req/sec. What would the logic be? What would the
: structure be? Where would you store this structure?
: 一家SF公司的面试题
| k***t
发帖数: 276 | 3 关于多机分布的，看到过一篇Paper的题目，好像有关。
"Cloud Control with Distributed Rate Limiting"。
不过我没读过这篇Paper：）谁有兴趣读一下，给版上的总结一下。
http://www.cs.ucla.edu/classes/cs217/SIG07Award.pdf
考虑？
【在 d********w 的大作中提到】
:
: 我想这个设计
: map>
: if (map.containsKay(developId))
: last_time, last_count = map.get(developId)
: if last_time != cur_time
: map.put(developId, new pair)
: else {
: last_count++;
: if (last_count > 10K)
|
|
