Discussion:
[squid-users] Delay pool not limiting each connection
Sekar Duraisamy
2018-09-17 13:15:43 UTC
Permalink
Hello All,

I have tried to limit the bandwidth for each requests with 5Mbps speed
with below.

delay_pools 1
delay_class 1 2
delay_access 1 allow all
delay_parameters 1 625000/625000 625000/625000


But it is limiting total bandwidth of the squid only 625000 and not
allowing the same bandwdith for each and every requests coming through
squid.

Squid version : Squid-3.5.27


Please help me if i missed anything.
Amos Jeffries
2018-09-17 15:02:10 UTC
Permalink
Post by Sekar Duraisamy
Hello All,
I have tried to limit the bandwidth for each requests with 5Mbps speed
with below.
delay_pools 1
delay_class 1 2
delay_access 1 allow all
delay_parameters 1 625000/625000 625000/625000
But it is limiting total bandwidth of the squid only 625000 and not
allowing the same bandwdith for each and every requests coming through
squid.
Yes. That behaviour is exactly what you configured above.

delay_parameters 1 aggregate individual


The "aggregate" is total bandwidth to be shared between individual
clients with each getting up to (but not exceeding) however much is
available.

By configuring aggregate and individual as exactly the same values you
have defined that any one client _could_ use the entire 625000 Bytes/sec
available. A second client only gets what that first one leaves unused.
A third only what those two leave unused, etc.

... which is exactly the "class 1" behaviour despite using a class 2
pool to do it.


To use a per-client limit, use a class 2 pools with an aggregate limit
of "none".


Also, no delay pool limits per-request. They are per-network,
per-subnet, per-client, per-user, or per-tag.

The closest you will get to a per-request limit is to have an
external_acl_type helper tag each request with a unique "tag="
identifier and use a class 5 pool.

However, please be aware that most traffic these days in HTTP/1.1 is
actually revalidation which only takes up a few hundred bytes, rarely
even a KB per-request. So having a Mbps per-request allocation is
pointless extra calculation and checking for a huge amount of requests.

Amos
Sekar Duraisamy
2018-09-18 03:07:51 UTC
Permalink
Anyone can help me on this?
Post by Sekar Duraisamy
Hello All,
I have tried to limit the bandwidth for each requests with 5Mbps speed
with below.
delay_pools 1
delay_class 1 2
delay_access 1 allow all
delay_parameters 1 625000/625000 625000/625000
But it is limiting total bandwidth of the squid only 625000 and not
allowing the same bandwdith for each and every requests coming through
squid.
Squid version : Squid-3.5.27
Please help me if i missed anything.
Arsalan Hussain
2018-09-18 03:53:48 UTC
Permalink
Dear Sekarit

you are using class-1 pool which has a single aggregate bucket to allocate
to all users and every one can use the maximum available FCFS basis.

you need to use class-2 delay pool has an aggregate bucket and 256
individual buckets which distributed among clients t keep them in limit
what you defined. (given below configuration as exampls)

acl bw_users src 192.168.1.0/24 # The acl defined for the Network
acl work_time time MTWHF 09:00-18:00
delay_pools 1 # Number of Pool
delay_class 1 2 # Defines the
class of pool for the Pool Number 1
delay_parametes 1 62500/62500 25000/25000 # each user has given an
average of 25000 bytes of bandwidth
delay_access 1 allow work_time # This is the access tag
which tie to the acl all and work_time.
Post by Sekar Duraisamy
Hello All,
I have tried to limit the bandwidth for each requests with 5Mbps speed
with below.
delay_pools 1
delay_class 1 2
delay_access 1 allow all
delay_parameters 1 625000/625000 625000/625000
But it is limiting total bandwidth of the squid only 625000 and not
allowing the same bandwidth for each and every requests coming through
squid.
Squid version : Squid-3.5.27
Please help me if i missed anything.
_______________________________________________
squid-users mailing list
http://lists.squid-cache.org/listinfo/squid-users
--
With Regards,


*Arsalan Hussain*
*Assistant Director, Networks & Information System*

*PRESTON UNIVERSITY*

*Hurt me with the truth, but never comfort me with a lie.*
Arsalan Hussain
2018-09-18 03:58:34 UTC
Permalink
you need to understand
delay_parameters 1 625000/625000 625000/625000

625000/625000 625000/625000( You defined 625000 max and also given 625000
to individual)

see this how you can limit each user bytes of bandwidht

delay_parametes 1 62500/62500 25000/25000 # each user has given
an average of 25000 bytes of bandwidth


w
Post by Arsalan Hussain
Dear Sekarit
you are using class-1 pool which has a single aggregate bucket to allocate
to all users and every one can use the maximum available FCFS basis.
you need to use class-2 delay pool has an aggregate bucket and 256
individual buckets which distributed among clients t keep them in limit
what you defined. (given below configuration as exampls)
acl bw_users src 192.168.1.0/24 # The acl defined for the Network
acl work_time time MTWHF 09:00-18:00
delay_pools 1 # Number of Pool
delay_class 1 2 # Defines the class of pool for the Pool Number 1
delay_parametes 1 62500/62500 25000/25000 # each user has given an average of 25000 bytes of bandwidth
delay_access 1 allow work_time # This is the access tag which tie to the acl all and work_time.
Post by Sekar Duraisamy
Hello All,
I have tried to limit the bandwidth for each requests with 5Mbps speed
with below.
delay_pools 1
delay_class 1 2
delay_access 1 allow all
delay_parameters 1 625000/625000 625000/625000
But it is limiting total bandwidth of the squid only 625000 and not
allowing the same bandwidth for each and every requests coming through
squid.
Squid version : Squid-3.5.27
Please help me if i missed anything.
_______________________________________________
squid-users mailing list
http://lists.squid-cache.org/listinfo/squid-users
--
With Regards,
*Arsalan Hussain*
*Assistant Director, Networks & Information System*
*PRESTON UNIVERSITY*
*Hurt me with the truth, but never comfort me with a lie.*
--
With Regards,


*Arsalan Hussain*
*Assistant Director, Networks & Information System*

*PRESTON UNIVERSITY*
Add: Plot: 85, Street No: 3, Sector H-8/1, Islamabad, Pakistan
Cell: +92-322-5018611
UAN: (51) 111-707-808 (Ext: 443)

*Hurt me with the truth, but never comfort me with a lie.*
Loading...