Transcript The Omega Testbed 1 - OpNeAR Lab

Sanjeev Lakshmanan, Dhruvish Shah, Kishore Gadhiraju, Marco Tacca, Andrea Fumagalli
Optical Networking Advanced Research(OpNeAR) Lab
The University of Texas at Dallas
IP
Centric Control Plane for
Survivable WDM Mesh
Networks
Overlay
Model with a
Distributed Architecture
Shared
Path Protection with
Differentiated Reliability
Graphical
User Interface
Generator
Receiver
The
– RWA & traffic generation
Linux PC’s act as controllers, run the
software
5
– RSVP & Protection switching
– Link Management Protocol for
detection of link and node failures
LMP
Synchronizer
7
Fast Ethernet links to form a Mesh
Topology
– Database
synchronization
An
electronic crossbar switch emulates a
crossconnect and is controlled by a device
driver
– For periodically discarding the
old packets from the network
Cacher
UNI
- User to Network Interface for
management and debugging
*
*Project is in collaboration with CpQD, Brazil
Resource Reservation protocol responsible for
dynamic end-end lightpath setup and release
hardware is emulated
A
delay element provides a configurable
delay on the ethernet links with a maximum
delay of 7.9ms
A
Incoming
connection requests enter any node,
which is then designated as the owner of that
connection
The
owner node runs the RWA, based on Shortest
Path routing and Least Loaded Wavelength
assignment, to find the working and protection paths
The
owner node first sets up the working path and
then reserves resources for the protection path
It initiates the Reservation procedure by sending a
LIGHTPATH_CREATE_TRY to the source node
Each
node in the path switches the ports of the
OXC, and forwards the message to the next node in
sequence.
In
case of a failure, the node sends a
LIGHTPATH_CREATE_FAIL to the previous node in
sequence and this message propagates all the way
back to the owner node.
In
case of a success on all the nodes in the path,
the destination node sends a SUCCESS message to
the source node
The
source node floods the SUCCESS message
Link
Management Protocol (LMP) runs
between adjacent nodes of every link
Each
node sends HELLO messages every
5 ms
Failure
is defined as the loss of 4
consecutive HELLO messages(20 ms)
Recovery
is defined as the reception of
100 consecutive HELLO messages(500 ms)