# US 2004/0064583 A1 · Configurable assignment of weights for efficient network routing

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2004/0064583 A1
- Document ID: US20040064583A1
- Status: Published application
- Kind: A1
- Application number: 10/259,336
- Filed: 2002-09-27
- Published: 2004-04-01
- Earliest claimed: 2002-09-27
- Assignees: Brocade Communications Systems, Inc.
- Inventors: Amod Dani; Asang Dani
- Web page: https://patentgrep.com/patent/US20040064583A1

## CPC

- H04L 67/1008: based on parameters of servers, e.g. available memory or workload (monitoring of computer activity G06F11/30)
- H04L 67/1001: for accessing one among a plurality of replicated servers
- H04L 67/101: based on network conditions
- H04L 67/1023: based on a hash applied to IP addresses or costs
- H04L: TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION (arrangements common to telegraphic and telephonic communication H04M)

## Abstract

A load balancing technique for use in a switch fabric balances network traffic from a plurality of end node devices based on the volume of traffic on the peripheral links, not simply the rated bandwidth of the peripheral links. In one embodiment, weight values, based on the volume of peripheral link traffic, are used to modify default cost values associated with the peripheral links. The modified default cost values are then used in the load balancing process. In another embodiment, the costs of the peripheral links are adjusted commensurate with the volume of peripheral link traffic.

## Claims

### Claim 1

A method of balancing network traffic in a network comprising a plurality of switches comprising end node ports coupled by peripheral links to end node devices, the method including:
(a) retrieving default cost values associated with said end node ports;
(b) modifying said default cost values to produce effective cost values associated with each of said end node ports; and
(c) balancing network traffic among said end node ports using said effective cost values.

### Claim 2 (depends on 1)

The method of claim 1 wherein, (b) includes multiplying said default cost values associated with said ports by weight values also associated with said ports.

### Claim 3 (depends on 2)

The method of claim 2 wherein said weight values are determined from the volume of network traffic on said peripheral links.

### Claim 4 (depends on 1)

The method of claim 1 further including programming a weight for each of said plurality of ports, said weight used in (b) to modify said default cost values.

### Claim 5 (depends on 4)

The method of claim 4 wherein said weight values are determined from the volume of network traffic on said peripheral links.

### Claim 6

A method of balancing network traffic in a network comprising a plurality of switches coupled via peripheral links to end node devices, the method including:
(a) programming cost values for a plurality of ports on a switch, said ports connected via peripheral links to end node devices, and said cost values programmed based on the volume of network traffic on said peripheral links; and
(b) balancing network traffic among said plurality of ports using said cost values.

### Claim 7 (depends on 6)

The method of claim 6 wherein (a) includes setting higher cost values associated with ports that have a volume of network traffic than is higher than for other ports.

### Claim 8

A method of balancing network traffic through a switch configured to be coupled to other switches via inter-switch links and to end node devices via peripheral links, the method including:
(a) assigning end node ports to ports connected to the inter-switch links based on traffic volume on the peripheral links.

### Claim 9

A network device, comprising:
a CPU that executes a load balancing process;
a memory coupled to said CPU, said memory containing storage for a weight buffer; and
a plurality of end node ports adapted to be connected to end node devices, each end node port having an associated default cost that relates to the rated bandwidth of a peripheral link connecting an end node device to the end node port;
wherein said weight buffer contains weight values associated with said end node ports; and
wherein said CPU modifies said default costs using said weight values to produce effective costs that are used to balance network traffic loads associated with said end node ports.

### Claim 10 (depends on 9)

The network device of claim 9 wherein said CPU modifies said default costs by multiplying said default costs by said weight values to produce said effective costs.

### Claim 11 (depends on 9)

The network device of claim 9 wherein said weight values are programmable.

### Claim 12 (depends on 11)

The network device of claim 11 wherein said weight values are determined from the amount of network traffic through said end node ports.

### Claim 13 (depends on 9)

The network device of claim 9 wherein said network device is a switch.

### Claim 14

A network device, comprising:
a CPU that executes a load balancing process; and
a plurality of end node ports adapted to be connected to end node devices via peripheral links, each end node port having an associated cost that is programmed commensurate with the volume of network traffic on the link connecting the end node device to the end node port;
wherein said CPU uses said costs to balance network traffic loads associated with said end node ports.

### Claim 15 (depends on 14)

The network device of claim 14 wherein said CPU sets higher cost values associated with end node ports that have a volume of network traffic than is higher than for other end node ports.

### Claim 16

A computer readable storage medium for storing an executable set of software instructions that are executable by a CPU, said software instructions being operable to balance network traffic through a plurality of end node ports on a network switch, comprising:
(a) a means for retrieving default cost values associated with the plurality of end node ports;
(b) a means for modifying said default cost values to produce effective cost values associated with each of said plurality of end node ports; and
(c) a means for balancing network traffic among said plurality of end node ports using said effective cost values.

### Claim 17 (depends on 16)

The storage medium of claim 16 wherein, (b) includes a means for multiplying said default cost values associated with said ports by weight values also associated with said port.

### Claim 18 (depends on 17)

The storage medium of claim 17 wherein said weight values determined from the volume of network on peripheral links adapted to couple to end node devices.

### Claim 19 (depends on 16)

The storage medium of claim 16 further including a means for programming a weight for each of said plurality of ports, said weight used in (b) to modify said default cost values.

### Claim 20 (depends on 19)

The method of claim 19 wherein said weight values are a function of the volume of network on said peripheral links.

### Claim 21

A computer readable storage medium for storing an executable set of software instructions that are executable by a CPU, said software instructions being operable to balance network traffic load from through a plurality of end node ports on a network switch, including:
(a) a means for programming cost values for a plurality of ports on a switch, said ports connected via peripheral links to end node devices, and said cost values programmed as a function of the volume of network traffic on said peripheral links; and
(b) a means for balancing network traffic among said plurality of ports using said cost values.
