Browse Prior Art Database

HANDLING I/O INTENSIVE WORKLOADS IN AFM CACHE THROUGH GATEWAY CONTROLLER

IP.com Disclosure Number: IPCOM000250459D
Publication Date: 2017-Jul-20
Document File: 1 page(s) / 54K

Publishing Venue

The IP.com Prior Art Database

Abstract

Introduce a new node designation in the AFM cache called the “gateway controller”. The gateway controller node can never be a gateway node itself. Business value: 1. Solving the problem using gateway controller helps to handle a highly active application in AFM without the need to add additional resources or call IBM for assistance for diagnosing the situation. 2. It would also warn the administrator thus giving enough time for any backup operations or control applications if that is possible. 3. Applications are not affected and are unaware of the problem and it is handled without any business impact. 4. High workload applications would be handled without any data integrity issues. 5. Lesser stringent planning during pre-deployment phases. Novelty of this approach: 1. Abstracting control through a gateway controller. 2. Having more than one gateway node to serve a cache fileset.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 53% of the total text.

HANDLING I/O INTENSIVE WORKLOADS IN AFM CACHE THROUGH GATEWAY CONTROLLER

Abstract

Introduce a new node designation in the AFM cache called the “gateway controller”. The gateway

controller node can never be a gateway node itself. 
 
 Business value: 1. Solving the problem using

gateway controller helps to handle a highly active application in AFM without the need to add additional resources or call IBM for assistance for diagnosing the situation. 2. It would also warn the administrator thus giving enough time for any backup operations or control applications if that is possible. 3.Applications are not affected and are unaware of the problem and it is handled without any business impact. 4. High workload applications would be handled without any data integrity issues. 5. Lesser stringent planning during pre-deployment phases. Key Features of this approach: 1. Abstracting control through a gateway

controller. 2. Having more than one gateway node to serve a cache fileset. 


Description

Introduce a new node designation in the AFM cache called the “gateway controller”. The gateway

controller node can never be a gateway node itself. 
 
 The role of the gateway controller would be:


 1. Monitor each gateway node's queue capacity at cluster level (which includes queue capacity of

all filesets in all filesystems owned by the gateway node). 
 
 2. Find a secondary gateway for a

fileset based on a gateway-choice algorithm, that finds a gateway node that has more room in queue

memory. 
 
 3. Re-direct further requests from filesets to a gateway node that has reached 95%

queue memory to a secondary gateway that has been selected based on a gateway-choice algorithm. Gateway controller updates the primary gateway node on the fileset’s record, so that application

nodes are aware. 
 
 4. After redirections of the overflowing requests to a secondary gateway node,

the gateway controller monitors both the original and the secondary gateway nodes at regular

intervals. 
 
 5. Schedules flushing of queues – by first having the primary gateway flush queue,

followed by queue flushed from the secondary gateway. This will keep up the ordering of operations

and to avoid any inconsistenci...