Simple Linux Proxy Controller/Forwarder

I need a Linux-based proxy controller/forwarded that will take requests from users, and use a proxy from our database to complete it. In simple terms: - User specifics {UserID}.[login to view URL]:{PortID} in their browser - "[login to view URL]" selects working proxy from DB and performs the request - "[login to view URL]" passes the response back to user's browser ============================ DB STRUCTURE ============================ TABLE: lists - id, proxy_ip, proxy_port, working, last_added, first_checked, first_working, last_checked, last_working, last_failed, total_fail_users, total_failreqs, real_ip, city, state, country, isp, isp_type, current_threads, max_threads TABLE: ports - id, userID, port, currentIP ([login to view URL]), rotate_seconds, current_threads, last_changed (last proxy change timestamp), timeout (after how many seconds of waiting should it consider current proxy failure), city, state, country, isp, isp_type TABLE: logs - id, userID, listID, portID, first_used, last_used ============================ WORKFLOW ============================ 1. User sets proxy criteria in the DB ("PortID") > ie. PortID "1" for UserID "a698j64p" is set to country=US, city=New York, isp=Comcast, rotate_seconds=300 >> In the above, user created PortID where they only want proxies from Comcast ISP in New York, and the proxy should rotate every 5 minutes 2. User will configure HTTP(s) or Socks4/5 proxy in their browser that points to: {UserID}.[login to view URL]:{PortID} > ie. [login to view URL] 3. If PortID for UserID has over X simultaneous threads/connections, deny it with an error (we want to limit the number of concurrent threads per port). Otherwise, continue to next step 4. "[login to view URL]" fetches PortID criteria from #1 and selects matching proxies in the DB > ie. SELECT [login to view URL],proxy_ip,proxy_port FROM lists LEFT JOIN logs ON [login to view URL] = [login to view URL] LEFT JOIN ports ON [login to view URL] = [login to view URL] WHERE [login to view URL] LIKE "comcast" AND lists.total_fail_users < 3 AND lists.total_fail_reqs < 3 AND lists.current_threads < lists.max_threads AND ([login to view URL] IS NULL OR logs.last_used < (NOW() - INTERVAL 60 MINUTE) AND ports.userID="a698j64p" AND ports.port=1 SELECT [login to view URL],[login to view URL],[login to view URL] FROM lists LEFT JOIN logs ON [login to view URL] = [login to view URL] LEFT JOIN ports ON lists.port=[login to view URL] WHERE [login to view URL] < 5 OR [login to view URL] IS NULL WHERE: - total_fail_users: # of users that encountered failure on the proxy - total_fail_reqs: #of requests (across all users) that encountered failure on the proxy - current_threads: To not overwhelm proxies, we might have to limit the number of concurrent threads connecting to it. This column will contain the current number of current threads (although I am not sure MySQL is the most optimal to use for this because of the massive amount of reads/writes required) - last_used: When the proxy was last used by the user 5. "[login to view URL]" forwards GET/POST request (with all headers) a) If proxy fails: - "[login to view URL]" updates total_fail_reqs++ and total_fail_users++ - Add "last_used" for proxy with current timestamp - Select different proxy and perform request again (user's connection should not drop) > if no remaining proxies, output error message and continue retrying every X seconds b) If proxy is a success, return response to user - Reset total_fail_reqs and total_fail_users - Add "last_used" for proxy with current timestamp Note: I am open to suggestions on any of the above. We are looking for the highest performance possible.