On 11/11/2013 5:04 p.m., Dr.x wrote: > Amos Jeffries-2 wrote >> On 2013-11-08 12:29, Dr.x wrote: >>> Amos Jeffries-2 wrote >>>> On 2013-11-08 11:26, Dr.x wrote: >>>>> . >>>>> >>>>> >>>>>> Select loops: >>>>>> * 1K/sec under the fast traffic period >>>>>> * relaying 3.5MB/sec >>>>>> >>>>>> * 7K/sec and 9K/sec in the periods you indicate as slow >>>>>> * relaying 4.7MB/sec >>>>>> >>>>>> => hints that Squid is looping once per packet or so. >>>>>> >>>>>> >>>>>> Amos >>>>> >>>>> something not being understood , >>>>> if u look at graph >>>>> u will note that "out" traffic is samller than "in" traffic !!!! >>>>> >>>>> >>>>> not understanding why !!!! >>>>> >>>> >>>> Think about what is "in" and "out" on that graph? >>>> Keep in mind that for each request Squid is handling two TCP >>>> connections, receiving and sending on both. Also performing HIT's on >>>> ~4% >>>> total HTTP traffic. >>>> >>>> >>>> Amos >>> >>> hi amos , im not talkign about t the difference in , out , >>> im wondering why the "in" is higher than "out" ??? >>> >>> shouldnt the "out" higher than "in" ( as a result of hit ration) ????? >>> >> >> That depends on what they are measuring. Which is why I asked. >> >> >>> i mean if i want to calculate wt im saving , i say (out-in)but in my >>> case >>> its in -ve !!!! >>> >> >> IF you measure "in" as being traffic on LAN interface and "out" as being >> traffic on WAN interface they could very well be negative. >> >> If you measure "in" as being packets into the box from any interface, >> and "out" as being packets leaving the Squid box. It could very well be >> *either* positive or negative. >> >> If you are measuring only one interface, they will again be *either* >> positive or negative depending on the interface. >> >> >> So ... what are "in" and "out" measuring *exactly* ?? >> >> >> Amos > > well , i must be missunderstanding something about the bw saving > calculation !!! > the question now is : > how i calculate the bw that im saving by squid ????? The full calculation is: bytes from client - bytes to server + bytes to client - bytes from server = bandwidth saving/loss. Current Squid only surface the metrics necessary to calculate that yourself in the "utilization" cachemgr report (the last section has totals) or SNMP counters (http://wiki.squid-cache.org/Features/Snmp#Squid_OIDs) under "Per-Protocol Statistics". Squid adds headers and sometimes chunked encoding overheads and strips some garbage from the headers as things go through. So even on a non-caching proxy there is a difference which may add up to a savings OR a loss. Amos
