[ 
https://issues.apache.org/jira/browse/HDDS-8211?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Mohanad Elsafty updated HDDS-8211:
----------------------------------
    Description: 
The target of QoS is to give all the users the opportunity to use S3Gateway 
even during periods of high load by limiting users making too many requests so 
the service becomes available for everyone.

 

The idea is taken from "{*}Apache Hadoop Fair Call Queue{*}"

 

!image.png|width=642,height=181!

 
 * Users will be making requests to S3Gateway which is intercepted by 
*ThrottlerFilter* that takes action whether to proceed with this request or 
reject it with the help of a scheduler.
 * At the moment there is one usable scheduler implementation 
*DecayRequestScheduler.*
 * *DecayRequestScheduler* keeps information about each user making requests 
and the availability of s3gateway handlers.
 * If all handlers are busy and some users are making too many requests it will 
start rejecting incoming requests with "{*}Service Unavailable{*}" response.
 * It also decays the number of user's requests count over the time.

  was:
The target of QoS is to give all the users the opportunity to use S3Gateway 
even during periods of high load by limiting users making too many requests so 
the service becomes available for everyone.

 

The idea is taken from "{*}Apache Hadoop Fair Call Queue{*}"

 

!image.png|width=486,height=137!
 * Users will be making requests to S3Gateway which is intercepted by 
*ThrottlerFilter* that takes action whether to proceed with this request or 
reject it with the help of a scheduler.
 * At the moment there is one usable scheduler implementation 
*DecayRequestScheduler.*
 * *DecayRequestScheduler* keeps information about each user making requests 
and the availability of s3gateway handlers.
 * If all handlers are busy and some users are making too many requests it will 
start rejecting incoming requests with "
Service Unavailable" response.
 * It also decays the number of user's requests count over the time.


> S3 QoS
> ------
>
>                 Key: HDDS-8211
>                 URL: https://issues.apache.org/jira/browse/HDDS-8211
>             Project: Apache Ozone
>          Issue Type: New Feature
>          Components: S3
>            Reporter: Mohanad Elsafty
>            Assignee: Mohanad Elsafty
>            Priority: Major
>         Attachments: image.png
>
>
> The target of QoS is to give all the users the opportunity to use S3Gateway 
> even during periods of high load by limiting users making too many requests 
> so the service becomes available for everyone.
>  
> The idea is taken from "{*}Apache Hadoop Fair Call Queue{*}"
>  
> !image.png|width=642,height=181!
>  
>  * Users will be making requests to S3Gateway which is intercepted by 
> *ThrottlerFilter* that takes action whether to proceed with this request or 
> reject it with the help of a scheduler.
>  * At the moment there is one usable scheduler implementation 
> *DecayRequestScheduler.*
>  * *DecayRequestScheduler* keeps information about each user making requests 
> and the availability of s3gateway handlers.
>  * If all handlers are busy and some users are making too many requests it 
> will start rejecting incoming requests with "{*}Service Unavailable{*}" 
> response.
>  * It also decays the number of user's requests count over the time.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to