metrics for network services
Download
Skip this Video
Download Presentation
Metrics for Network Services

Loading in 2 Seconds...

play fullscreen
1 / 20

Metrics for Network Services - PowerPoint PPT Presentation


  • 719 Views
  • Uploaded on

Metrics for Network Services. CIT 443. Service Metrics. Timeliness Capacity Quality. Service Metrics. Measurable Valid Accurate Timely. Why use metrics?.

loader
I am the owner, or an agent authorized to act on behalf of the owner, of the copyrighted work described.
capcha
Download Presentation

PowerPoint Slideshow about 'Metrics for Network Services' - lynne


An Image/Link below is provided (as is) to download presentation

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.


- - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - -
Presentation Transcript
service metrics
Service Metrics
  • Timeliness
  • Capacity
  • Quality
service metrics3
Service Metrics
  • Measurable
  • Valid
  • Accurate
  • Timely
why use metrics
Why use metrics?

Metrics are key indicators that can be used to identify and understand the behaviors, successes, and failures of specific functions, processes, or systems

timeliness
Timeliness
  • Response Time
    • Amount of time between a request for service and the receipt of the subsequent reply
    • Users vs Computers
      • Users are impatient
      • Interactivity
      • Computers can be extremely patient
timeliness6
Timeliness
  • Latency/delay
    • Time between data being ready to send and the actual transmission of that data
    • Occurs at multiple points in a network
      • Why?
    • Often expressed as the total latency for a path between two nodes
    • Latency is a fact of life
timeliness causes of latency
Timeliness – Causes of Latency
  • Physics
    • The speed of the light is the speed of light…
  • Signaling
  • Switching Delay
  • Queue Depths
timeliness delay variations
Timeliness- Delay Variations
  • Amount of variation in delay
    • Jitter
  • Has a large negative effect on streaming technologies/protocols
    • If it is predictable, then it can potentially be buffered
  • A good target: 2% (or less)
  • Frame sizing
capacity
Capacity
  • Bandwidth vs Throughput
  • Ideally, throughput = capacity
    • Highly dependent on underlying technologies in use
    • 2nd order equation
capacity10
Capacity
  • Utilization
    • % of total availability in use at any given time
    • Optimal Utilization
      • Maximum average utilization before the network is considered saturated.
      • Depend on underlying technology
        • Ethernet vs Token Ring
capacity11
Capacity
  • Offered Load
    • Sum of the data the network has available to send at any given time
  • Data Rates
    • Peak Data Rate
    • Sustained Data Rate
    • Minimum Data Rate
capacity estimating data rates
Capacity - Estimating Data Rates
  • Distributed Computing
    • 103 seconds (with data size 105 MB)
  • Web Transactions
    • 10 seconds (with data size 102 MB)
  • Database Transactions
    • 10 seconds (with data size 102 MB)
  • Teleconferencing (multicasting)
    • 1 second (with data size .5 MB)
capacity13
Capacity
  • Scalability
    • Network Growth
    • Long Term vs Short Term
    • “Bottleneck Management”
    • 80/20 rule
quality
Quality
  • Accuracy
  • Efficiency
  • Availability
  • Reliability
accuracy
Accuracy
  • The percentage of traffic that is error free
  • Accuracy for WAN links is usually described as being under or over the threshold bit error rate (BER)
    • Analog lines BER=1 in 105
    • Digital lines (copper) BER=1 in 106
      • Vs 109 – 1010 for LAN
    • Digital lines (fiber) BER=1 in 1011
  • Accuracy for LAN links is typically described in terms of frames error rate
efficiency
Efficiency
  • Amount of useful data transmitted compared to total number of bits transferred
  • Inverse percentage of overhead
    • Smaller headers=more efficiency
    • Larger frames/packets/cells=more efficiency
      • Assuming error free transmission
  • Dependent upon technology
availability
Availability
  • What percentage of time is the network up and available to meet the business needs of customers compared to the total time the network is designed to be operational
  • Expressed as a percentage
    • Or number of nines
  • Cost of downtime
    • Mission critical systems cannot be down without significant loss of revenue opportunity costs
  • No mention of when the network is available
    • Most networks have critical times that the system much be functional
  • High performance
    • Availability >= 99.9%
reliability
Reliability
  • How often the network fails and how long it takes to repair it
  • MTBF
    • Mean Time Between Failures
    • High performance MTBF >=8000 hours
  • MTTR
    • Mean Time To Repair
    • High performance MTTR <= 2 hours
  • Both of these parameters have individual component/link bases and end-to-end bases
    • End-to-end is the product of the individual bases
    • .999 * .999 * .999 = .997
types of service
Types of Service
  • Best effort
    • As much bandwidth as we can get
  • Deterministic Services
    • Estimate of required bandwidth, but no means of guaranteeing bandwidth
  • Guaranteed Services
    • A means of enforcing bandwidth is implemented
    • As bandwidth approaches upper boundaries policing takes place to limit bandwidth consumption
ad