You are on page 1of 27

Fundamental Switching Types

• Circuit Switching
• Virtual Circuit Switching
• Datagram Switching
• Implications for Signaling, Routing, Path
Computation, and Restoration
– MPLS and GMPLS control planes

Page - 1 © Grotto Networking 2004


Differences in Switching Types

• Is connection set up required?


• Is statistical multiplexing possible?
• What are the QoS measures? How is
bandwidth allocated?
• How much work is needed to provide QoS
guarantees?
• How can reliability/protection/restoration be
provided and what are the trade offs?

Page - 2 © Grotto Networking 2004


Forwarding at each switch

• Datagram (e.g., IP)


– Based on complete destination address within the packet.
Any valid destination must be forwarded correctly.
• Virtual Circuits (e.g., MPLS, ATM, Frame Relay)
– Based only on a label with the packet header. Only
packets whose “virtual circuit” has been set up ahead of
time must be forwarded correctly.
• Circuits (not packets)
– Based implicitly on either time slot or wavelength. No
forwarding information needed in data. Only those
circuits whose path has been set up ahead of time must be
forwarded correctly.

Page - 3 © Grotto Networking 2004


Example Network

– Datagram, Virtual Circuits, or Circuits


– Switches 1-5, Hosts A-J

Page - 4 © Grotto Networking 2004


Datagram Forwarding Example
Switch #1 Switch #2 Switch #3 Switch #4 Switch #5
Dest Port Dest Port Dest Port Dest Port Dest Port
A 1 A 2 A 1 A 1 A 1
B 2 B 2 B 1 B 1 B 1
C 3 C 1 C 1 C 3 C 1
D 3 D 3 D 1 D 3 D 1
E 4 E 2 E 2 E 1 E 2
F 4 F 2 F 4 F 1 F 2
G 4 G 4 G 3 G 2 G 2
H 4 H 4 H 3 H 4 H 2
I 3 I 4 I 3 I 3 I 3
J 3 J 4 J 3 J 3 J 4
Graph of our
example network
I with switch ports
I and hosts shown
I I I
I

Page - 5 © Grotto Networking 2004


Virtual Circuit forwarding Example
• Connections
– Host A to Host J, Host B to Host C, Host E to Host I,
Host D to Host H, and Host A to Host G

Page - 6 © Grotto Networking 2004


Virtual Circuit Forwarding

– Packets are forwarded based on a label in the header


– Labels are not destination addresses, usually much
shorter
– Labels need to be unique on a link but not in a network,
i.e., we can reuse labels on each link.
– Switch forwarding tables consist of a map between
(input port, packet label) to (output port, new packet
label)
– Table entry for each virtual circuit rather than for each
destination (the datagram case)
– Technologies: MPLS, Frame Relay, ATM, X.25

Page - 7 © Grotto Networking 2004


VC Forwarding Table Example
Switch #1 Switch #2 Switch #3
In Port In Label Out Port Out Label In Port In Label Out Port Out Label In Port In Label Out Port Out Label
1 2 3 5 2 5 4 1 1 1 3 3
2 1 3 1 2 1 1 1 2 1 3 1
1 1 4 1 3 6 4 3

Switch #4 Switch #5
In Port In Label Out Port Out Label In Port In Label Out Port Out Label
1 3 2 5 1 1 4 2
1 1 3 1 1 3 2 1
3 1 4 1 6 2 1 3 1

3
3
1

1
Page - 8 © Grotto Networking 2004
“Real” Circuit Forwarding

• No more packets
• Bit streams are distinguished by port and
– Time slots in the TDM case
– Wavelength in the WDM case
– Frequency in the FDM case
• Switching independent of bit stream contents
• TDM example (same connections as VC case)
– Host A to Host J, Host B to Host C, Host E to Host I,
Host D to Host H, and Host A to Host G

Page - 9 © Grotto Networking 2004


“Real” Circuit Tables Example
Switch #1 Switch #2 Switch #3
In Port In Slot Out Port Out Slot In Port In Slot Out Port Out Slot In Port In Slot Out Port Out Slot
1 2 3 5 2 5 4 1 1 1 3 3
2 1 3 1 2 1 1 1 2 1 3 1
1 1 4 1 3 6 4 3
Switch #5
Switch #4 In Port In Slot Out Port Out Slot
In Port In Slot Out Port Out Slot 1 1 4 2
1 3 2 5 1 3 2 1
1 1 3 1 2 1 3 1
3 1 4 1

Page - 10 © Grotto Networking 2004


Time Division Multiplexing
TDM Path
Path Path
Multiplex Section MS MS
Regenerator
RS RS RS RS
Section
TDM TDM de-
Multiplexor multiplexor
Regenerator Regenerator
(3R) #1 (3R) #2

= Optical Fiber

= Regenerator section overhead


= Multiplex section (line) overhead
= User traffic (path layer)
= Unused time slots

Page - 11 © Grotto Networking 2004


Real Circuits and Virtual Circuits

• Virtual Circuits
– Packet based, label (not destination address) in packet
header
– Doesn’t always consume bandwidth, i.e., traffic can be
bursty
• Real Circuits
– No packets raw bit stream, implicit label with either
time slot or wavelength
– Is always consuming a fixed bandwidth, easy to keep
track of bandwidth but not necessarily the most
efficient utilization of link capacity.

Page - 12 © Grotto Networking 2004


QoS with Real Circuits

• Bandwidth
– Hard bandwidth guarantees are given by default
(even if you don’t want them).
• Delay
– Very little delay variation. Most delay
attributable to propagation. Switching delays in
most circuit switches is minimal.
• Bit Error Rate
– Is the primary “signal quality measure”

Page - 13 © Grotto Networking 2004


QoS with Virtual Circuits

• Bandwidth
– Is by default shared with other users. Effort required to
make guarantees. Very good statistical multiplexing
gain can be obtained.
• Delay
– In addition to propagation and switch processing delay
we now have queueing induced delays
– Queueing delays: can be quite large, can be quite
variable
– By default no guarantees made
• Dropped/Errored Packets
– Packets can be errored (bits errors), or dropped due to
buffer overflows.
Page - 14 © Grotto Networking 2004
Protection/Restoration
• Failure detection
– Most circuit technologies have very fast built-in failure
detection.
– For packet technologies this hasn’t been the case but
new work, e.g., BFD at IETF is underway.
• Alternative Routes
– Alternate routes for circuit consume bandwidth or must
be set up on the fly costing time.
– Alternate routes for virtual circuits do not consume
bandwidth until they are used, hence can be set up
ahead of time.
– Alternate routes can not be preconfigured for datagram
networks and all switches (routers) must recalculate
routing tables based on link failure info.
Page - 15 © Grotto Networking 2004
Forwarding Tables

• All switching types use them


• Datagram Forwarding Tables
– Need to account for all destinations no matter who’s
communicating at any given time.
• Circuit and Virtual Circuit Forwarding Tables
– Entry for each circuit or VC that traverses a particular
switch.
– Note that if there are N host and they all want to talk
to each other at exactly the same time then the network
will need to support N(N-1) circuits or VCs.

Page - 16 © Grotto Networking 2004


Scaling Forwarding Tables

– Modern networks like the Internet and Telephone


networks consists of 100 of Millions or more hosts how
can we keep our routing tables under control?
• Datagram Tables
– We route based on networks and groups of networks.
Addresses are given out accordingly. This allows the
aggregation of destination addresses.
• Circuit Tables
– We multiplex circuits onto larger and larger trunks in a
hierarchy. Switches generally only work at a couple
levels of the hierarchy. Example a switch working with
SONET OC48 links (2.5Gbps) will switch with
50Mbps granularity but not 64kbps granularity!
Page - 17 © Grotto Networking 2004
Setting up the Routing Tables

• Finding Paths from Source to Destination


– How do we choose our “route”?
– Algorithms
– Protocols
• Datagram Routing
– Must make sure that the tables are consistent so
we don’t get datagram loops.
• “Real” Circuit Routing
– Need to have enough bandwidth available on
the links to support the circuits.
Page - 18 © Grotto Networking 2004
Differences Between Optical Network
Routing and IP Routing

• IP routing
– Per hop forwarding of datagrams based on destination
IP address
– Every router must have exactly the same network
topology information (links, nodes, and link wts.)
– Every router must run exactly the same path
computation algorithm
– Failure to insure these last two requirements can result
in routing loops and “black holes”

Page - 19 © Grotto Networking 2004


Differences Between Optical Network
Routing and IP Routing...
• Optical routing
– Circuits are source routed; no loops possible
– No standardization of path computation required
– Okay for information to be slightly out of date, e.g.,
available capacity information; worst case “crank-back”
of connection
– Unless restoration action is taken based on link state
updates, routing is not service impacting in transport
domain

Page - 20 © Grotto Networking 2004


What is GMPLS?

• GMPLS = Generalized MPLS


– Refers to adaptation of MPLS control plane for
the control of other technologies
– Includes signaling and routing mechanisms
developed for MPLS traffic engineering
– GMPLS protocols developed under IETF
– Previously called “MPλS”

Page - 21 © Grotto Networking 2004


What is MPLS?

• MPLS = “Multi-Protocol Label Switching”

• A virtual circuit form of packet switching


such as frame relay or ATM but with a
more IP centric control plane and built in IP
“adaptation”.

Page - 22 © Grotto Networking 2004


MPLS and IP

• A combine IP router / MPLS switch assigns IP


packets to MPLS flows (virtual circuits)
– This process is known as “classification” and can be
very simple or very complex depending upon the
context.
– This box is known as a Label Edge Router (LER)
• The IP packet header is not touched or looked at
while in the MPLS network
– The LSR (label switched routers) only switch based on
MPLS labels.

Page - 23 © Grotto Networking 2004


An example…

Ingress Label switching & Egress router


router adds packet forwarding removes label
label to packet

IP
Unlabeled
IP 20
Packet arrives
0
IP 1

IP

Autonomous
system boundary

Page - 24 © Grotto Networking 2004


Label Switched Path (LSP)

Label switched path

A Label Switched Path is like a pipe or tunnel to IP packets.


However its just another term for a virtual circuit. While traveling
on a label switched path, forwarding is based on the label only,
not on destination IP address in packet.
Page - 25 © Grotto Networking 2004
Controlling LSP Set-Up: Explicit
Routing
12.0.0.1

10.1.1.1 10.1.1.2 10.1.1.3

10.1.1.5 10.1.1.6

Explicit route
10.1.1.4 10.1.1.7
10.1.1.7 strict
10.1.1.6 strict
10.1.1.5 strict
10.1.1.2 strict
Strict hop
POP LSP takes direct route to 10.1.1.7
10.1.1.1 strict

Page - 26 Similar procedure can be used


© Grotto for optical
Networking 2004 connection set-up.
“Generalized” MPLS

• Virtual Circuits ! Real Circuits


– From real labels in MPLS to “virtual labels” in GMPLS
• “Labels” in GMPLS
– TDM where time slots are the implicit labels (e.g.,
SONET)
– FDM where frequencies (or λs) are the implicit labels
(e.g., WDM)
– Space-division multiplexing where port numbers are
the implicit labels (e.g., OXCs)
• Generalized labels used in MPLS messaging =
Generalized MPLS

Page - 27 © Grotto Networking 2004

You might also like