CCDA 640-864 Summary Notes – Chapter 6 – Day 14

Exam Topic 3: WAN Design Methodology

Use PPDIOO Methodology. Key concepts:

  • Analyze the network requirements
    • Application type
    • Traffic volume
    • Traffic patterns
  • Characterize existing network
  • Design topology
    • chould be flexible and have room for growth
  • Implement High Availability. Consider:
    • Response time
    • Throughput
    • Reliability
  • Application requirements
    • Data File transfer
      • Reasonable response time (time between the cient user request and the response from the server host)
      • High Throughput (throughput – measure of data transferred from one host to another in a given amount of time)
      • Medium packet loss tolerance
      • Reliability: Reasonable downtime (Reliability – measure of a given application’s availability to its users + whether the service is performing as it should)
    • Interactive Data Applications
      • < 1sec response time
      • Low Throughput
      • Low packet loss tolerance
      • Reliability: Low downtime
    • Real Time Voice
      • RTT < 250 ms (delay + jitter)
      • Low throughput
      • Low packet loss tolerance
      • Reliability: Low downtime
    • Real Time Video
      • Min delay and jitter
      • High throughput
      • Medium packet loss tolerance
      • Reliability: Minimum downtime

Bandwidth

Bandwidth considerations for different mediums

  • Copper
    • < 2 Mbps
      • Serial
      • ISDN
      • FrameRelay
      • TDM
      • ASDL
    • 2 – 45 Mbps
      • Frame Relay
      • Ethernet
      • ADSL
      • Cable T3
    • 45 – 100 Mbps
      • Fast Ethernet
    • 100Mbps – 10 Gbps
      • GE
      • 10GE (10GBaseCX4)
  • Fiber
    • < 2 Mbps -NA
    • 2 – 45 Mbps
      • Ethernet
    • 45 – 100 Mbps
      • FE
      • ATM
    • 100Mbps – 10 Gbps
      • GE
      • 10GE
      • ATM
      • SONET/SDH
      • POS
      • dark fiber
  • Wireless
    • < 2 Mbps
      • 802.11b
    • 2 – 45 Mbps
      • 802.11b
      • wireless WAN (varies)
    • 45 – 100 Mbps
      • 802.11a/g
    • 100 Mbps – 10 Gbps
      • 802.11n

When Link utilization

  • reaches 50 to 60 % consider increasing bandwidth and closely monitor
  • 75% – immediate attention is required to avoid congestion , packet loss

WAN Link Categories

Private

  • Use:
    • Connects distant LANs
  • Cost:
    • Expensive. Owner must buy and maintain
  • Advantages:
    • High security
    • Transmission Quality
  • Examples:
    • Metro Ethernet using dark fiber

Leased

  • Use:
    • Connects distant LANs
  • Cost:
    • High Cost. Equipment is leased or private
  • Advantages:
    • Maintenance is done by provider
    • Dedicated bandwidth
  • Examples:
    • TDM
    • SONET

Shared

  • Use:
    • Shared circuit or packet-switched WAN
  • Cost:
    • Fair
    • Leased bandwidth
    • Leased or private equipment
  • Advantages:
    • Provider responsible for maintenance
    • Shared network for multiple sites
  • Examples:
    • MPLS
    • Frame Relay

Cost Types:

  • Fixed
    • network equipment
    • circuit provisioning
    • network management tools
  • Recurring
    • service provider monthly WAN service fees
    • maintenance costs of the WAN
    • network operations personnel

Bandwidth Optimization using QoS

QoS ensures that the most critical traffic gets the best treatment and available bandwidth in times of congestion

QoS Mechanisms:

Queuing – the process of buffering traffic that is used by network appliances when the rate of incoming traffic is greater than the rate at which they are processing the traffic

Classification – the process of identifying and marking traffic. The traffic is assigned priority according to type. Example: NBAR which is used for deep packet content classification I order to identify the type of traffic. It can identify traffic at the application layer and also traffic that does not use standard ports. Committed Access Rate CAR also supports classification by the use of an access list.

Congestion Management – mechanism that is used to handle traffic when congestion is prevalent using different queuing techniques. There are 2 types of queues in a router, the hardware and software queue. Hardware queue always handles traffic in a first in first come basis. The software queue is used only when there is congestion In order to schedule traffic using different queuing techniques, police or shape the traffic before sending it to the hardware one.

Priority queue: traffic is passed to either the high, low, default or medium traffic output queues. The high priority queue must be emptied before the other queues and that Mingus lead to queue starvation.

Custom queuing is a legacy queuing system fairer than PQ. The traffic has upto 16 output queues that have customisable byte sizes.  All queues are serviced at one time or the other

Weighted fair Queuing WFQ. Traffic is separated into flows. High bandwidth flows have precedence over low bandwidth flows. Both flows are assigned weights. Default for interfaces with 2 Mbps or less speed

Class Based weighted fair queuing CBWFQ ( Modular QoS CLI MQC ) This is enhanced WFQ. It is modular in nature. Traffic classes are user defined and have corresponding queues. After matching the traffic, it’s characteristics can be modified. When congestion occurs, only the guaranteed characteristics are provided

Low Latency Queuing ( LLQ) = PQ + CBWFQ. PQ provides for prioritisation of low latency traffic. The queue has a maximum threshold so as not to starve other queues.

Traffic Shaping, and Policing are used for congestion avoidance.
Shaping is used to slow down the rate at which traffic goes out of an interface. The provider also can avoid dropping traffic that exceeds the CAR. It eliminates bottlenecks. It uses a to ken bucket technique. Shaping buffers packets.
Policing (CAR) uses the leaky bucket technique to limit the amount of traffic coming into an interface. Policing can be configured to drop packets.

Link efficiency is used on slow speed links to reduce the amount of latency and or jitter. The mechanisms include the following:
1) link fragmentation and interleaving – breaks larger packets and inserts smaller ones between them
2) multilink PPP – bonds multiple links between two nodes In order to increase the bandwidth
3) RTP header compression – compresses the header from 40 bytes to 2-5 bytes on slower links.

TCP window size- this is the number of frames that can be received without an acknowledgement. It can be adjusted to minimise delay. If the window size is not adjusted, retransmission can occur.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s