eG Monitoring
 

Measures reported by PKTSSPartStTest

A partition manages bandwidth for a traffic class' aggregate flows, so that all of the flows for the class are controlled together as one.

Partitions can be used to:

  • Protect mission-critical traffic by guaranteeing that a traffic class always gets a defined amount of bandwidth

  • Limit aggressive, non-critical traffic by allowing that traffic class to consume only a defined amount of bandwidth

  • Divide capacity

  • Assign bandwidth dynamically to users

  • Oversubscribe your link

There are two types of partitions that can be created: static or dynamic.

A static partition manages bandwidth for all flows within a particular traffic class. Static partitions can be fixed or burstable.

  • A fixed partition allows an aggregate traffic class to use a defined amount of bandwidth, if needed. A fixed partition not only ensures that a specific amount of bandwidth will be available, but it also limits traffic to that same level.

  • A burstable partition allows an aggregate traffic class to use a defined amount of bandwidth, and also allows that traffic class to access additional unused bandwidth, if needed. You can put a cap on a burstable partition, allowing it to access upto a maximum amount of bandwidth, or you can allow a burstable partition potentially to consume all available bandwidth.

In situations where administrators may want to apply bandwidth limits to individual users, they can establish dynamic sub partitions for the traffic class. A dynamic partition carves up a static partition's bandwidth, creating sub partitions on the fly for new users. Subpartitions are children of a static partition.

Partitions protect traffic by guaranteeing a defined amount of bandwidth for mission-critical traffic classes. If the traffic to a partition is gradually increasing or if suddenly administrators notice too much of data/packet drops on a partition, then, the traffic flow to the traffic class may be disrupted. This may lead to attacks on the servers/devices that are connected to the load balancers. To avoid such abnormalities, administrators may often need to monitor the partitions round the clock. The PKTSSPartStTest test helps administrators in this regard!

This test auto-discovers the partitions that are available in the target PacketShaper S-Series load balancer and reports how well data/packets were transmitted by each partition. Administrators can capture packets that were dropped and figure out how frequently data/packets were retransmitted. This way, problematic partitions on which frequent data/packet drops were noticed can be isolated and administrators can start investigating on the real issues leading to such data/packet drops.

Outputs of the test : One set of the results for each partition configured on the target PacketShaper Load Balancer S-Series device that is being monitored

The measures made by this test are as follows:

Measurement Description Measurement Unit Interpretation
dataCount Indicates the amount of data transmitted by this partition during the last measurement period. MB The detailed diagnosis of this measure lists the direction of the WAN Link and the rate at which data was transmitted over the partition.

Comparing the value of this measure across partitions will help administrators identify the partition that is transmitting maximum amount of data.
reXDataCount Indicates the amount of data that was retransmitted by this partition during the last measurement period. MB If large amount of data takes too long to be successfully retransmitted, you may have to figure out what is causing repeated retransmission failures and fix it before packet loss occurs.
packetCount Indicates the number of packets that were transmitted by this partition during the last measurement period. Number Comparing the value of this measure across partitions will help administrators identify the partition that is transmitting maximum number of packets.
dataPacketCount Indicates the number of data packets that were transmitted over TCP by this partition during the last measurement period. Number  
reXDataPktCount Indicates the number of data packets that were retransmitted over TCP by this partition during the last measurement period. Number If TCP packets take too long to be successfully retransmitted, you may have to figure out the real reason behind repeated retransmission failures and fix it before packet loss occurs.
lateDropktCnt Indicates the number of packets that were dropped by this partition during the last measurement period. Number Packets are dropped if they are not transmitted from the queue after a wait time of eight seconds or more.

Ideally, the value of this measure should be zero.
lateDropDataCnt Indicates the amount of data that that was dropped by this partition during the last measurement period. Number Data can be dropped if the data is not transmitted from the queue after a wait time of eight seconds or more.

Ideally, the value of this measure should be zero.