Professional Documents
Culture Documents
Lower Layers
Function:
Application Route packets end-to-end on a
Presentation network, through multiple hops
Session Key challenge:
Transport How to represent addresses
Network
How to route packets
Data Link
Scalability
Physical Convergence
How Do You Get IPs?
4
Simplex
One direction
e.g. Television
Half duplex
Either direction, but only one way at a time
e.g. police radio
Full duplex
Both directions at the same time
e.g. telephone
Time domain concepts
9
Analog signal
Various in a smooth way over time
Digital signal
Maintains a constant level then changes to another constant level
Periodic signal
Pattern repeated over time
Aperiodic signal
Pattern not repeated over time
Analogue & Digital Signals
10
Periodic Signals
11
Wavelength, Spectrum & Bandwidth
12
Wavelength
Distance occupied by one cycle
Distance between two points of corresponding phase in two consecutive cycles
Spectrum
range of frequencies contained in signal
Absolute bandwidth
width of spectrum
Effective bandwidth
Often just bandwidth
Narrow band of frequencies containing most of the energy
Data Rate and Bandwidth
13
Cheaper
Less susceptible to noise
Greater attenuation
Pulses become rounded and smaller
Leads to loss of information
Router Design
Router Architecture
15
Data Plane
Moving the data, i.e., the packets
How packets get forwarded
Data plane only needs to know the “FIB” (Forwarding Information Base)
Smaller, less information, etc.
Simplifies line cards vs the network processor
Control Plane
How routing protocols establish routes/etc.
Control plane must remember lots of routing info (BGP tables, etc.)
Generic Router Architecture
16
Header Processing
Data Hdr Data Hdr
Lookup Update Queue
IP Address Header Packet
Interfaces
Input/output of packets
Switching fabric
Moving packets from input to output
Software
Routing
Packet processing
Scheduling
Etc.
Summary of Routing Functionality
18
Route processor
Routing
Installing forwarding tables
Management
Line cards
Packet processing and classification
Packet forwarding
Cumbersome to configure
Cannot adapt to addition of new links or nodes
Cannot adapt to link or node failures
Cannot easily handle multiple paths to a destination
Does not scale to large networks
Solution is to use Dynamic Routing
Desirable Characteristics of Dynamic Routing
30
Scalability
Robustness
Simplicity
Rapid convergence
downtime
Packets don’t get to where they are supposed to go
Black holes (packets “disappear”)
Routing Loops (packets go back and fore between the same devices)
Occurs when there is a change in status of router or the links
Interior Gateway Protocols
32
RIPv2 is classless
has improvements over RIPv1
is not widely used in the Internet industry
Onlyuse is at the internet edge, between dial aggregation devices
which can only speak RIPv2 and the next layer of the network
IGRP/EIGRP
34
Forwarding = moving
packets between interfaces
according to the “directions”
Forwarding decisions:
Destination address
class of service (fair queuing, precedence, others)
local requirements (packet filtering)
Routing Protocols
40
Proactive protocols
Traditional distributed shortest-path protocols
Based on periodic updates. High routing overhead
Reactive Routing
41
query(0)
Disadvantages:
3 high flood-search overhead with
reply(0)
query(0) high route acquisition latency
query(0)
2
4
query(0)
reply(0) query(0)
5
Reactive Routing – Source initiated
43
Cooperative nodes
Relatively small network diameter (5-10
hops)
Detectable packet error
Start Route no
Buffer Route
Discovery found?
packet
Protocol
yes
Continue
normal
wait
Route Packet
Discovery in
finished
buffer? no Send packet
to next-hop
45 CNSM6114 – Network Programming done
Route Discovery: At an intermediate node
no
Host’s
address yes Discard
already in route
patrial request
route
Append no
myAddr to no
partial route myAdd
r=targe
t
yes
Store <src,id> in
list Send route
reply packet
Broadcast packet
46 CNSM6114 – Network Programming done
AODV Routing Protocol
47
S E
F
A
C
G D
B
S E
F
A
C
G D
B
S E
F
A
C
G D
B
S E
F
A
C
G D
B
S E
F
A
C
G D
B
S E
F
A
C
G D
B
What is it?
Goal: sending a packet from A to B
Applications
TCP UDP
IP
MPLS MPS
PPP FR ATM Ethernet DWDM
Physical
LSR - Routers that support MPLS are called Label Switch Router
LER - LSR at the edge of the network is called Label Edge Router
(a.k.a Edge LSR)
Ingress LER is responsible for adding labels to unlabeled IP packets.
Egress LER is responsible for removing the labels.
Label Switch Path (LSP) – the path defined by the labels through LSRs
between two LERs.
Label Forwarding Information Base (LFIB) – a forwarding table
(mapping) between labels to outgoing interfaces.
Forward Equivalent Class (FEC) – All IP packets follow the same path
on the MPLS network and receive the same treatment at each node.
MPLS Applications
58
Traffic Engineering
Virtual Private Network
sites over a carrier’s network. Each site has its own private IP
address space.
Different VPNs may use the same IP address space.
Circuit switching
Packet switching
Circuit Switching
64
Establish
Transfer
Disconnect
Must have switching capacity and channel capacity to
establish connection
Must have intelligence to work out routing
Circuit Switching - Applications
65
Inefficient
Channel capacity dedicated for duration of connection
If no data, capacity wasted
Set up (connection) takes time
Once connected, transfer is transparent
Blocking
A network is unable to connect stations because all paths are in
use
A blocking network allows this
Used on voice systems
Short duration calls
Non-blocking
Permits all stations to connect (in pairs) at once
Used for some data connections
Packet Switching Principles
67
Line efficiency
Single node to node link can be shared by many packets over time
Packets queued and transmitted as fast as possible
Data rate conversion
Each station connects to the local node at its own speed
Nodes buffer data if required to equalize rates
Packets are accepted even when network is busy
Delivery may slow down
Priorities can be used
Datagram
70
missing packets
Datagram
Diagram
71
Quality Of Service
Best Effort vs. QoS
73
Best Effort:
You get a link to the Internet with at most B bits/sec.
If you don’t like it, switch to another provider.
Worse-case
Provide bandwidth/delay/jitter guarantee to every packet
E.g., “hard real time”
Average-case
Provide bandwidth/delay/jitter guarantee over many
packets
Statistical in nature
Quality of service issues
75
Flow specification
Flow spec: traffic characteristics, QoS requirements (delay, jitter, bandwidth)
Routing
Routing traffic to best meet demand
Resource reservation
End-host signaling to network QoS resource requirements
Admission control
Limiting number of reservations
Packet scheduling
Packet by packet scheduling (fairness, delay)
RSVP addresses reservation
Worse-case : Guaranteed Services
76
Service contract
Network to client: guarantee a deterministic upper
bound on delay for each packet in a session
Client to network: the session does not send more
than it specifies
Algorithm support
Admission control based on worst-case analysis
Per flow classification/scheduling at routers
Average-case: Controlled Load Service
77
Service contract:
Network to client: Average delay, jitter, bandwidth, e.g., makes
network appear as an unloaded, best effort network with
bandwidth and delay
Client to network: the session does not send more than it
specifies
Algorithm Support
Admission control based on measurement of aggregates
Scheduling for aggregate possible
Role of RSVP in the Architecture
78
At each hop
Consult admission control and policy module
Set up admission state or informs the requester of
failure
Network Processor
Network Processor
80
Function
An intelligent NIC
Higher Performance
Specialized network processing engines
Multiple processing elements
Low Latency
Intelligence
Network level without going to main processor
Modularity
Taking the processing load off GPP
NP handles the network
GPP handles the application
Application
84
Forwarding (bridging/routing)
Protocol Conversion
In-system data movement (DMA+)
Encapsulation/Decapsulation to fabric/backplane/custom devices
Cell/packet conversion (SAR’ing)
L4-L7 applications; content and/or flow-based
Security and Traffic Engineering
Firewall, Encryption (IPSEC, SSL), Compression
Rate shaping, QoS/CoS
Intrusion Detection (IDS) and RMON
Particularly challenging due to processing many state elements in parallel, unlike most
other networking apps which are more likely single-path per packet/cell
Acceleration Techniques
85
Dealing with shared state among parallel units requires some form of locking
and/or sequential consistency control which can eat some of the benefit of
parallelism
Caveat; more parallel activity increases memory contention, thus latency
Latency Hiding via Hardware Multi-Threading
88
Sometimes specialized hardware is the best way to get the required speed for
certain functions
Many NP’s provide a fast path to external coproc’s; sometimes slave
devices, sometime masters.
Variety of functions
Encryption and Key Management
Lookups, CAMs, Ternary CAMs
Classification
RegEx string searches (often on reassembled frames)
Statistics gathering
A Typical NP Architecture
90
Network General
Physical Network Coproc
Purpose
(i.e. GbE) Interface DMA/Buffer Interface
Processor
Internal BUS
Memory DMA/BUS
Memory Coproc
Interface Interface
Interfaces
Network (Myrinet – High BW/Low Latency)
BUS Interface
Myrinet Cards
92
Embedded Processor
I/O Processor
Peer-to-peer
Network Processor
PCI Interface
One to the Main BUS
Other to the Network Interface
Similar to Myrinet LANai
Further development leading into IXA?
Intel IXA
95
Current Routers
Involve general purpose CPUs
Lots of ASICs (Application Specific Integrated Circuits ).
The ASICs are necessary to keep up with the quantity and rate of the
network traffic.
The StrongARM Core
Replace the general purpose CPUs
Microengines
Replace the bulk of the ASICs
Actually inherited IXA when they bought Digital.
Intel IXP1200 NP
96
The End!