On 06/20/2018 01:44 PM, Michael Pro wrote: > Is it possible to limit the speed of receiving the file to the client > depending on the file size? > Let's say, > file =<1MB client download with 400Kbit/s > file 1MB...5MB client donload with 4Mbit/s > file >5MB - no speed limit For HTTP responses with a Content-Length header, you can create poor man's equivalent of a missing response_body_size ACL. For example: acl smallerThanTenBytes rep_header Content-Length ^[0-9]$ With an ACL like that, you can direct responses to the right pool. IIRC, pool assignment may happen multiple times during the lifetime of a transaction so you need to be careful to deal with assignments that are done before there is a response. Delay pools are not my area of expertise but others on the list can help you with the details. For HTTP responses without a Content-Length header, you can use an adaptation service (ICAP or eCAP) to tag transactions that start to exceed 1MB and 5MB boundaries. A "note" ACL can check for that tag, but I do not think Squid delay pools re-check pool ACLs after the response body bytes start to flow. If I am right, then you would not be able to (re)route response body bytes to the right delay pool for responses without a Content-Length header. You could use an adaptation service itself to slow responses without Content-Length down (partially duplicating Squid's delay pools capabilities), but the precision of such an approach may be too low due to Squid internal buffering. YMMV. For HTTP CONNECT tunnels (i.e., HTTPS traffic) even the adaptation tricks would not work because adaptation services do not get CONNECT tunnel bytes (yet?). You would have to bump the tunnel to get access to the HTTP responses inside (and bumping opens another huge Pandora box). HTH, Alex. _______________________________________________ squid-users mailing list squid-users@xxxxxxxxxxxxxxxxxxxxx http://lists.squid-cache.org/listinfo/squid-users
