Last Week

1
Last Week

• Hierarchical WAN/LAN Design
• Three Layer
• Two Layer
• One Layer

2
(No Transcript)
3
A System Approach to Network Design

• Requirements analysis
• Flow Analysis
• Logical Design
• Technology choices
• Interconnection mechanisms
• Network Management and security
• Physical Design
• Addressing and Routing

4
(No Transcript)
5
Logical Design Interconnection Mechanisms

• Background
• Shared Medium (no interconnection)
• Switching and Routing
• Switching
• Routing
• Hybrid Switching/Routing Mechanisms
• Applying Interconnection Mechanisms to the Design
• Case Study

6
Logical Design Network Management and Security

• Integrating Network Management and Security into
  design
• Defining Network Management
• Designing with Manageable Resources
• Network Management Architecture
• Security
• Security Mechanisms
• Security Examples
• Network Management and Security Plans
• Case Study

7
Logical Design Interconnection Mechanisms

• Background
• Technology choices made
• Need interconnection mechanisms to connect these
  choices together
• Shared medium (no interconnection)
• Bridging/Switching (Ethernet/FDDI/Token Ring),
  Asynchronous Transfer Mode (ATM) switching, frame
  switching, Switched Multi-megabit Data Service
  (SMDS)
• Network layer routing

8
Logical Design Interconnection Mechanisms

• Background
• Interconnection mechanisms
• ()
• Hybrid mechanisms
• LAN Emulation (LANE)
• Multiprotocol over ATM (MPOA)
• Next-hop resolution protocol (NHRP)
• Current trend integrating network layer and data
  link layer functionality

9
Logical Design Interconnection Mechanisms

• Background
• Factors in evaluating interconnection mechanisms
• Scalability
• Optimising flow models
• Providing external connectivity to the network
• Applying support services to the design

10
Logical Design Interconnection Mechanisms

• Shared Medium (no interconnection)
• A large version of an area network
• Combine networks through physical media or hubs,
  and repeaters
• A common IP addressing structure
• All hosts on the network directly accessible by
  each other

11
Logical Design Interconnection Mechanisms
shared medium
12
Logical Design Interconnection Mechanisms
shared medium

• Simplest solution
• Lack of scalability
• More users/hosts/applications leads to growing
  overhead (multicast, address resolution, network
  layer functions)
• Performance degrades
• Network becomes unusable
• Conclusion Hierarchy is essential

13
Logical Design Interconnection Mechanisms
shared medium

• Planned hierarchy
• Isolate and hide networks and their traffic from
  one another
• Add new networks or
• Segment existing ones

14
(No Transcript)
15
Logical Design Interconnection Mechanisms
shared medium

• To meet growth expectations
• Scaling
• Select appropriate technologies
• Plan upgrades in the technology
• Plan hierarchy

16
Logical Design Interconnection Mechanisms
shared medium

• When do we need to introduce hierarchy into the
  shared medium network
• When the maximum data rates (MDRs) for the
  expected traffic flows on a shared-medium network
  approach 80 of the capacity for the technology
  used in that network
• When the sustained data rates (SDRs) for the
  traffic on a shared medium network approach 60
  of the capacity for the technology used in that
  network
• Need for redundancy

17
Logical Design Interconnection Mechanisms
shared medium

• 4. When external connectivity is needed
• IP routing or protocol gateway for Internet
  connectivity
• 5. For certain flow models
• Client-server
• Collaborative computing
• Distributed computing
• 6. For security, administration and network
  management

18
Logical Design Interconnection Mechanisms

• Switching and Routing
• Methods of moving information between network (NL
  or DLL)
• Switches create connections for the duration of
  the information flow their switching
• Routers oriented towards connectionless transfer
  (connections may occur at transport layer), have
  more buffers than switches

19
Logical Design Interconnection Mechanisms

• Switching and Routing
• Switching Generally at Data Link Layer
• Link layer Ethernets, Token Rings, FDDIs, HiPPIs
• Single IP network
• Routing Generally at Network (IP) Layer
• Separates interconnected networks via IP
• Switches
• Cut through
• Fragment free (collisions are not switched)
• Fast forward switching
• Store and forward

20
Logical Design Interconnection Mechanisms

• Switching and Routing
• Routers Store and forward
• Routers are more tolerant of traffic bursts than
  switches
• Routers introduce longer and more variable delays
• Routers have much more network management,
  security and user interface capabilities than
  switches

21
Logical Design Interconnection Mechanisms

• Switching
• Bridge when you can, route when you must not
  quiet!
• Relatively simple and usually cheap way to
  achieve scalability and to optimise sum flow
  models
• Trade offs Simplicity and speed (switch) versus
  complexity and function (router)

22
Logical Design Interconnection Mechanisms

• Switching
• Multiple connections through the switch at the
  same time at media speeds
• When a flow model suggests multiple simultaneous
  flows, switches can be used to optimise the
  network
• Distributed computing
• Client server computing
• Cooperative computing
• For WAN connectivity routers are used

23
Distributed Computing
24
Distributed Computing
25
Routers in a WAN Distributed Computing Environment
26
ATM Switching

• Backbone?
• Desktop?
• 1992 Threat to Frame Relay, Ethernet, Token
  Ring, FDDI
• Today Network backbones WANs

27
ATM Switching

• Can provide some support to end to end services
• Complex
• High level of capacity
• Path for performance updates (ATM over SONET)

28
ATM Switching

• High overhead
• Cell size 48 bytes data 5 bytes header 53
  bytes (9.4 overhead)
• for segmentation and reassembly up to 8 bytes
  for ATM Adaptation Layers (24.5 overhead)
• LANE interconnects technologies such as Ethernet
  Token Ring with ATM
• Provides address resolution, broadcast
  capabilities, connectionless service
• Looks like a bridged network, use router at NL

29
ATM Switching

• LANE
• Use if you have a large base of existing Ethernet
  and Token Ring networks minimize cost
• Complex, reduced manageability
• Improved performance over Ethernet or Token Ring
• Can be used as high performance backbone

30
ATM Switching

• With ATM Backbone
• Direct ATM connection or
• Fast/Gigabit Ethernet, HiPPI, or Fibre Channel
• IP over ATM
• More straight forward
• Routing used to interconnect
• Trade-off not taking full advantage of ATM,
  strange interconnectivity patterns, simple and
  straight forward interconnection at IP layer

31
Routing

• Already discussed

32
Hybrid Switching/Routing Mechanisms

• Using the benefits of switching and routing
• ATM options
• LAN Emulation (LANE)
• Multiprotocol over ATM (MPOA)
• Next-hop resolution protocol (NHRP)
• Service switching

33
Next-hop resolution protocol (NHRP)

• One method for diverging from the standard IP
  routing model to optimise paths in the network

34
Multiprotocol over ATM (MPOA)

• Applies NHRP to LANE
• Combines switching and routing

35
Service switching

• Uses a criteria to determine whether to route or
  switch traffic
• A hybrid switching/routing environment

36
Evaluation Criteria for Interconnection Mechanisms
37
Evaluation Criteria for Interconnection Mechanisms
38
Applying Interconnection Mechanisms to Design

• Importance of interconnections
• Hierarchy
• Redundancy

39
Hierarchy

• Play an important role
• Indicate consolidation points in the network
• Likely to show where multiple technologies
  interconnect
• Metrics for measuring hierarchy
• Concentration of flows
• Technology connections at or within an area
• Number of networks/areas interconnected

40
Degree of hierarchy
41
Redundancy

• Number of alternate paths in a network
• Provides reliability
• Alternate paths can provide a variety of degrees
  of redundancy in the network
• Low redundancy path
• Medium redundancy path
• High redundancy path

42
Low redundancy path

• First, an alternate path may not be immediately
  available when the primary path is disabled
• Configuration may be required (significant delay
  - minutes)
• Alternate path reduces performance
• E.g. T1 (1.544 Mb/s, 250 ms delay) satellite link
  backs up T3 (45 Mb/s, 10 ms delay) link.

43
Medium redundancy path

• An alternate path is readily available when the
  primary path is disabled
• Immediately available or
• Short wait (ms to second)
• Alternate path reduces performance
• E.g. T1 (1.544 Mb/s, 10 ms delay) link backs up
  T3 (45 Mb/s, 10 ms delay) link.

44
High redundancy path

• One or more alternate paths are readily available
  when the primary path is disabled
• Alternate path offers equal performance
• E.g. T3 (45 Mb/s, 10 ms delay) link backs up T3
  (45 Mb/s, 10 ms delay) link.

45
Degrees of redundancy
46
Hierarchy Redundancy

• The guidelines for hierarchy and redundancy in
  the network are useful in
• Developing interconnection mechanisms
• Developing network management, security, and
  addressing and routing strategies for our design.

47
Logical Design Network Management and Security

• Integrating Network Management and Security into
  Design
• Defining Network Management
• Designing with Manageable Resources
• Network Management Architecture
• Security
• Security Mechanisms
• Security Examples
• Network Management and Security Plans

48
Integrating Network Management and Security into
Design

• Defining and characterising management for a
  network design
• How to plan for
• Monitoring
• Configuring and
• Troubleshooting the network
• Network management protocols and instrumentation
  requirements
• Architectural aspects of applying a network
  management to the design

49
Defining Network Management

• Tasks
• Monitoring
• Configuring
• Troubleshooting
• Accounting
• Planning
• Etc.

50
Defining Network Management

• Categories of tasks
• Monitoring for event notification
• Monitoring for metrics and planning
• Configuration of network parameters
• Troubleshooting the network

51
Next week

• Defining Network Management
• Designing with Manageable Resources
• Network Management Architecture
• Security
• Security Mechanisms
• Security Examples
• Network Management and Security Plans

52
Next Week

• Introduction to Queuing theory