Белая книга для Cisco Cisco Nexus 5010 Switch
© 2014 Cisco and/or its affiliates. All rights reserved. This document is Cisco Public Information.
Page 4 of 7
Figure 1. Example of a Microburst in the Network
In Figure 1, three servers are connected to a switch. Two servers are sending and one server is receiving the
traffic. The arrows from the servers on the left to the switch represent the traffic flow, and the graph next to the
servers shows the traffic in each flow. The arrow on the right represents traffic flow from the egress port on the
switch to the destination server. The graph above the switch shows traffic in the flow from the switch to the
destination server, which is sum of the two signals from the source servers. At the marked point, two ingress
signals create a microburst at the destination, and if the switch does not have enough resources to buffer the part
of the destination traffic shown in orange, that traffic will be dropped.
The Cisco Nexus 5600 platform and Cisco Nexus 6000 Series Switches introduce a microburst monitoring feature
that observes this behavior and notifies the network administrator about its occurrence. The network administrator
can then use the Switched Port Analyzer (SPAN) feature to collect traffic and mitigate the microburst.
Microburst Monitoring Feature Concept
Microburst monitoring on the Cisco Nexus 5600 platform and Cisco Nexus 6000 Series Switches can detect
microbursts on a per-port basis in both the ingress and egress traffic directions. In the case in which a server with a
bursty application is connected to a port, microbursts typically occur in the ingress direction. When servers
simultaneously access a storage array, microbursts occur in the egress direction on the port going to the storage
array.
Microburst monitoring detects microbursts when a specific number of packets or a specific amount of data (in
bytes) is exceeded during a defined time interval. The switch counts every occurrence and tallies the information in
a table. This table allows the network administrator to quickly correlate application behavior with a server and
network port. When this information is used in combination with logging information, the exact time of the event can
be known.