Browse Prior Art Database

Use fluid dynamics for effective data locating on network

IP.com Disclosure Number: IPCOM000015539D
Original Publication Date: 2002-Apr-01
Included in the Prior Art Database: 2003-Jun-20
Document File: 1 page(s) / 43K

Publishing Venue

IBM

Abstract

Growth of network makes effective locating of data on network an important concern for system administrators. When a system becomes large, it becomes difficult to determine what is a most effective locating of data apri-ori. This invention provides a way to determine data location in a systematic way and dynamically. Amount of data (a large file which can be split into parts or a file system or a group of file systems) is represented by a fluid which is spread over a network. A system is modeled by a way described in the following. Map network topology to a 2D plane preserving distances in network. Capacity of storage node corresponds to area of a region on the plane. Data moves according to an equation of motion of the fluid. 2 Action is S (dP/dt) mP2 Vt(P) U(z) P P(t;z) is density of data at z (x,y) of the moment. It should be [integral of P over the plane] 1. Kinetic term appears because rapid movement of massive data costs network load. 2 The term mP raises the system energy when data is concentrated on a point one particular server. V [integral of (P x |z-z'|) over the plane] is caused by access to a file from a point z' in the certain period of the past. Data located far from access point raised system energy. U(z) represents potential barriers. A long distance hop in the network causes this term that suppress movement of data beyond it.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 93% of the total text.

Page 1 of 1

Use fluid dynamics for effective data locating on network

Growth of network makes effective locating of data on network an important concern for system administrators. When a system becomes large, it becomes difficult to determine what is a most effective locating of data apri-ori. This invention provides a way to determine data location in a systematic way and dynamically.

Amount of data (a large file which can be split into parts or a file system or a group of file systems) is represented by a fluid which is spread over a network. A system is modeled by a way described in the following.

Map network topology to a 2D plane preserving distances in network. Capacity of storage node corresponds to area of a region on the plane. Data moves according to an equation of motion of the fluid. 2Action is S = (dP/dt) + mP2 + Vt(P) + U(z) P = P(t;z) is density of data at z = (x,y) of the moment. It should be [integral of P over the plane] = 1. Kinetic term appears because rapid movement of massive data costs network load.

2The term mP raises the system energy when data is concentrated on a point = one particular server. V = [integral of (P x |z-z'|) over the plane] is caused by access to a file from a point z' in the certain period of the past. Data located far from access point raised system energy. U(z) represents potential barriers. A long distance hop in the network causes this term that suppress movement of data beyond it.

By solving this equation of motion, it becomes...