communication 5 A 'mess i

US 20050047420A1

(19) United States (12) Patent Application Publication (10) Pub. No.: US 2005/0047420 A1 (43) Pub. Date:

Tanabe et al. (54) MOBILE IPV6 NETWORK HAVING

Mar. 3, 2005

Publication Classi?cation

MULTIPLE HOME AGENTS AND METHOD OF LOAD BALANCE

(51)

Int. Cl? ................................................... .. H04L 12/56

(52)

Us. 01. ...................................................... ..370/395.52

(76) Inventors: Shiro Tanabe, Hidaka (JP); Xiao Long

Huang, Beijing (CN); Hui Deng, Beijing (CN); Kai Zhang, Beijing (CN); Zhisheng Niu, Beijing (CN)

(57)

Correspondence Address: ANTONELLI, TERRY, STOUT & KRAUS,

In a mobile IPv6 network having multiple distributed regres sion home agents and a load balance method for the multiple

LLP 1300 NORTH SEVENTEENTH STREET SUITE 1800

regression home agents, the network comprises a plurality of mobile subnets connected to each other through an Internet. Each mobile subnet comprises an access router, a plurality

ARLINGTON, VA 22209-9889 (US)

(21) Appl. No.:

10/880,650

(22) Filed:

Jul. 1, 2004

(30)

ABSTRACT

of mobile nodes, and a plurality of regression agents. The regression agents are arranged in a distributed topology structure. The regression agents exchange information With each other by performing a broadcast of traffic load infor

Foreign Application Priority Data

mation (table) among the regression agents. Further, each of the regression agents has a traffic load table to perform the

Jul. 1, 2003

load balance operation accordingly.

(CN) .................................... .. 03145741.X

6_\ .

.

communication node

jntemet

I

/

I

ll

, f.n

i regresion regresion :5: I

:

L

angle router

’___

5A

‘mess i

E

router

:

_

,r‘egfesion

i‘

mobile node

i :

foreign mobile

sub-network 1

‘?i‘~\i> router access R3 :

=

|

_

*|

locol mobile sub—network regression network I

_|

i

agent HA1 0 en HA/2/l’ ................. . agent O HAn L

:

:

.

:

:

'

j

l mobile "ji -------node __

router access R2 %

i

i

i.

I

i

i

fore'gn mobne

i

.

L

sub-network 2

J

8

: -

.

i

Patent Application Publication Mar. 3, 2005 Sheet 1 0f 5

US 2005/0047420 A1

__

.|i .|rl in

M m EE2:who:239:26:3525

m85 E32:8I/l

:N {3, a5:gm:_-1J.

_ n _ \ _ |.

N .

:. w

m 5.56: _"@202\2965 $32 m .n __ n_J__L632 M m 5%: F258w _SO_ 2505 {Same? 05258

m O iliwwoc m W 0 \$320:39.:

.ol #

JwI\\i.11:I1. 1. _ _ _ .. . ..

L"c58528.l:2“@.F3l%r5&l2Aur\n.\\\E:E»?Iu.2:E08"oC”888:. .

\ i .m

Patent Application Publication Mar. 3, 2005 Sheet 3 0f 5

US 2005/0047420 A1

FIG. 3 regression agent lP

Load

regression agent 2 regression agent i

Lood=600Kbps Load=300Kbps

regression agent 12 LIGHT regression agent 7 UGHT regression agent x

UGHT

registered mobile node number

4 8 20 4 ' 3

regresion regresion

regresion

agent HA1 0 ent HA2

agent HAn

access

G b """"""""""""" Q a} router R3

l

regression network color repoesents load size of regression agent

Patent Application Publication Mar. 3, 2005 Sheet 4 0f 5

US 2005/0047420 A1

FIG. 4 timer exceeds binding time 8]

determine whether the load is larger than a threshold or not?

yes

no

S2 Detemine whether there is a

Detemine whether there registered

"LIGHT'regression agent or not? \no

yes

Randomly select 54 one of the

"LIGH l"regression

‘

53 mobile in all "LIGHT" regression .

'

agent is top 10% or not‘?

Detemine whether in all no

"LlGHT'regression is top 10% or not?

55 yes

' no

agent,and return;

' Randomly select one of bottoml0%

Randomly select one

of bottoml0% regression

56 agents in the non

"LlCHT"regression agents, -

and return;

Perform no handoff

S8 regression agents in

S7 operation of the

all"LlGHl"regression

regression agent,

agents,and return.

and return;

'

Patent Application Publication Mar. 3, 2005 Sheet 5 0f 5

US 2005/0047420 A1

(a.N825E8