Transcript Tree Configuration in Bridged IEEE1394 Bus Network Subrata Banerjee PHILIPS Research Briarcliff P1394.1 WG Meeting March 19-20, 1998 PHILIPS Research.

Tree Configuration in Bridged
IEEE1394 Bus Network
Subrata Banerjee
PHILIPS Research Briarcliff
P1394.1 WG Meeting
March 19-20, 1998
PHILIPS Research
Problem Statement
 Bridges of different capabilities


Bandwidth
Iso Delay
 Bridges may introduce unacceptable loops
 Path between any two bridges

“Minimum bottleneck” route
N1
C1
C2
N2
 Typically bridge capacity lower than bus capacity
PHILIPS
Subrata Banerjee
[email protected]
Graph Theory Result
 Maximum Spanning Tree guarantees
Minimum Bottleneck Route
Between every pair of nodes
 Distributed maximum spanning tree algorithm
C1
C2
PHILIPS
Subrata Banerjee
[email protected]
How to Choose Between Two Bridges?
 Bridge Capability Parameters

Bridge Bandwidth

Bridge Iso_Delay

Bridge Vendor ID
= max. of 2 portal vendor IDs

Bridge Node ID
= max. of 2 portal vendor IDs
 Proposed Organization
Bridge BW (13)
Resv (3) Iso_delay (8)
Resv (8)
Chip ID (40)
Vendor ID (24) (scrambled)
PHILIPS
Subrata Banerjee
[email protected]
An Example of Tree Conf. Algorithm (1)
57
57
65
65
52
21
55
62
40
48
48
31
58
27
38
38
PHILIPS
58
27
36
Subrata Banerjee
[email protected]
36
34
An Example of Tree Conf. Algorithm (2)
 BP Broadcasts
65
BP57
65
52
21
BP40
BP31
58
27
48
48
31
55
BP52
40
BP62
BP48
58
27
BP34
BP27
BP38
BP58
BP36
38
38
PHILIPS
BP55
BP21
BP65
62
57
57
36
Subrata Banerjee
[email protected]
36
34
An Example of Tree Conf. Algorithm (3)
 Fragment Roots
65
57
57
BP Count=3
BP Count=3
65
52
21
62
65, L1
BP Count=2
40
62, L1
31
BP Count=4
38
27
BP Count=2
58
58, L1
48, L1
38
PHILIPS
58
27
48
48
57, L1
55
36
Subrata Banerjee
[email protected]
36
34
An Example of Tree Conf. Algorithm (4)
 Find Best Neighbor
65
65
57
57
52
21
65, L1
62
40
62, L1
31
38
58
58, L1
48, L1
38
PHILIPS
58
27
Submit
48
27
48
57, L1
55
36
Subrata Banerjee
[email protected]
36
34
An Example of Tree Conf. Algorithm (5)
 Join 1
57
57
65
65
52
21
65, L1
62
40
62, L1
31
38
58
27
58, L1
48, L1
38
PHILIPS
58
27
48
48
57, L1
55
36
Subrata Banerjee
[email protected]
36
34
An Example of Tree Conf. Algorithm (6)
 Join 1
57
57
65
65
52
21
57, L1
55
62
65, L2
40
48
48
31
38
58
27
58, L1
48, L1
38
PHILIPS
58
27
36
Subrata Banerjee
[email protected]
36
34
An Example of Tree Conf. Algorithm (7)
 Find new best neighbor
65
65
57
57
52
21
57, L1
55
62
65, L2
40
48
48
31
38
58
27
58, L1
48, L1
38
PHILIPS
58
27
36
Subrata Banerjee
[email protected]
36
34
An Example of Tree Conf. Algorithm (8)
 Join 2
57
57
65
65
52
21
57, L1
55
62
65, L2
40
48
48
31
38
58
27
58, L1
48, L1
38
PHILIPS
58
27
36
Subrata Banerjee
[email protected]
36
34
An Example of Tree Conf. Algorithm (9)
 Join 2
57
57
65
65
52
21
55
62
65, L2
40
48
48
31
38
58
58
27
48, L1
38
PHILIPS
27 58, L2
36
Subrata Banerjee
[email protected]
36
34
An Example of Tree Conf. Algorithm (10)
 Join 3
57
57
65
65
52
21
55
62
65, L2
40
48
48
31
38
58
58
27
48, L1
38
PHILIPS
27 58, L2
36
Subrata Banerjee
[email protected]
36
34
An Example of Tree Conf. Algorithm (11)
 Join 3
57
57
65
65
52
21
55
62
65, L3
40
48
48
31
27 58, L2
58
27
38
38
PHILIPS
58
36
Subrata Banerjee
[email protected]
36
34
An Example of Tree Conf. Algorithm (12)
 Join 3
57
57
65
65
52
21
55
62
65, L3
31
40
48
48
38
PHILIPS
27 58, L2
58
58
27
36
Subrata Banerjee
[email protected]
36
34
An Example of Tree Conf. Algorithm (13)
 Join 4
57
57
65
65
52
21
55
62
40
31
48
48
38
PHILIPS
27 58, L2
58
58
27
36
Subrata Banerjee
[email protected]
36
34
An Example of Tree Conf. Algorithm (14)
 Join 4
57
57
65
65
52
21
55
62
40
31
58
48
48
38
PHILIPS
58
27
36
Subrata Banerjee
[email protected]
34
An Example of Tree Conf. Algorithm (15)
 Join 4
57
65
65
57
52
55
62
40
48
48
PHILIPS
Subrata Banerjee
[email protected]
58
58
Rules of the Game
 Fragment A can submit to Fragment B iff
LevelA  LevelB
 Bridge Capability values are unique.
C1
 No loops possible
Lx
Lx
Ly
Lx
C1 > C2 > C3 > C1
Lx > Lz > Ly > Lx
C2
C3
Lx
Lz
PHILIPS
Subrata Banerjee
[email protected]
Example of Selected Commands
 “I am a BP”

data = Unique Bridge Capabilities (UBC)
 “Report UBC”

data = best UBC from all children
 “Connect”

data = Fragment Level
 “Update”

data = Fragment UID, Fragment Level
PHILIPS
Subrata Banerjee
[email protected]
Once the Tree Topology is Identified ...
 Assign

Bus Ids

Routing Bounds

Portal_Control.rte fields
 Select Net Cycle Timer
 Assign Portal_Control.clk fields
 Details?
PHILIPS
Subrata Banerjee
[email protected]