Preconfigured Graphs
This section describes all the graphs available in the Statistics:Reports
window. (Not available on the PacketShaper 900 Lite, 1200, or 1400 Lite models.) Click a graph name
below to see a description and a sample graph. You can also download a
printer-ready PDF file
of all the graphs.
These graphs are grouped into the following basic categories,
or types of graphs:
Graphs for Analyzing Bandwidth Usage
| Graph Name |
When to use it |
| Class
Utilization |
To analyze a class' average bandwidth usage |
| Class
Utilization with Peaks |
To see if a class is frequently hitting a capacity limit |
| Dynamic
Partition Usage |
To determine whether the size of a dynamic partition is appropriate |
| Link Utilization |
To determine whether the link usage varies a lot and to see what your average capacity needs are
To see whether the link size has changed over the interval (if you choose to display the link size when creating the graph) |
| Link
Utilization with Peaks |
To examine the utilization of an inbound or outbound link
To compare a link's average and peak bandwidth usage with the link
size so that you can determine whether the link size is sufficient (if you choose to display the link size when creating the graph) |
| Partition Utilization |
To examine a partition's average bandwidth usage
To compare a partition's average bandwidth usage with the partition's size so that you can determine whether the size is appropriate (if you choose to display the partition size when creating the graph) |
| Partition
Utilization with Peaks |
To determine the average and peak bandwidth usage of a partition
To compare a partition's average and peak bandwidth usage with the partition's size (if you choose to display the partition size when creating the graph) |
Graphs to Analyze Efficiency
| Graph Name |
When to use it |
| Bytes
Transmitted |
To determine the actual volume (such as number of gigabytes) that
has been transmitted in a time period for a specific link, partition,
or class |
| Guaranteed
Rate Failures |
To see the number of times PacketWise was unable to provide the
bandwidth guaranteed by the class' policy; allows you to analyze the
effectiveness of a policy |
| Network
Efficiency |
To see what percentage of your network traffic is devoted to packet
retransmissions. Actually, this graph shows the percentage of packets
that are not retransmissions, so make sure you interpret this
graph correctly. For example, if network efficiency is 90%, then 10%
of the packets are retransmissions. |
| Packet
Size Distribution |
To analyze the distribution of packet sizes on an Inbound or Outbound
link |
| Packets
Transmitted |
To determine the cause of a network slowdown. Compare the Packets
Transmitted chart to the corresponding class or link utilization chart
to determine if a link or class is being flooded by small packets.
The flooding of small packets reduces the rate at which full-size
packets get transmitted, which could slow down the network. |
Graphs for Analyzing Top Ten
| Graph Name |
When to use it |
| Top
10 Partitions |
To quickly see which partitions are consuming the most bandwidth
on the network |
| Top
10 Classes |
To quickly see which classes are consuming the most bandwidth on
the network |
|
Top
10 Children Classes
|
To see the biggest consumers of bandwidth for a class' direct descendant
children |
Graphs for Analyzing Response Time
| Graph Name |
When to use it |
| Average
Transaction Size |
To view information about transaction size as part of interpreting
RTM data. For example, you can compare a Network Delay graph and an
Average Transaction Size graph for the same time period to get a better
picture of what is causing a delay. |
| Network
Delay |
To analyze the portion of transaction time that is attributable
to the network. Since network delay is affected by transaction size,
the Normalized
Network Delay graph could be more useful when there is high variability
in transaction size. |
| Network
Delay Distribution |
To see the breakdown of network delays for a class, partition, or
link |
| Normalized
Network Delay |
To analyze response-time data accounting for variation in transaction
sizes. |
| Packet
Round-Trip Time |
To track the history of network speed, independent of the size of
the transactions |
| Server
Delay |
To analyze the portion of transaction time that is attributable
to the server |
| Server
Delay Distribution |
To see the breakdown of server delays for a class, partition, or
link |
| Service
Level Compliance |
When you've set a total delay threshold and service level threshold,
you can see how a class is performing with respect to the thresholds;
the graph shows the percentage of transactions whose delays are in
an acceptable range |
| Transaction
Delay |
To determine the source (server or network) of any significant delays
that a traffic class experiences; to compare actual response times
to the total delay threshold |
| Transaction
Delay Distribution |
To get a more complete picture of transaction delays; this graph
shows the number of transactions whose total delay falls into various
transaction-time buckets; for example, you can see that most transactions
experience .1 ms or .25 ms total delay and only a few experience 1
ms. |
Graphs for Analyzing TCP Connections
| Graph Name |
When to use it |
| Connection
Retransmissions |
To identify poorly tuned servers whose retransmissions are wasting
bandwidth; to show applications that could benefit from traffic shaping
or rate control |
| TCP
Connections Initiated |
To view the rate at which new connections are opened |
| TCP
Health |
To view a comprehensive picture of TCP connections for a class,
partition, or link. Use this graph to identify abnormal network conditions,
analyze why a server isn't performing properly, or determine whether
a DoS attack occurred. |
Graphs for Analyzing Compression
Use these graphs to analyze how much compression your PacketShaper
is doing. You can analyze compression per class, partition, or the entire
link.
|
