PassivePacketLoss
From GEANT2-JRA1 Wiki
Contents |
Packetloss - Monitoring of packet loss for real traffic
Description
Packetloss is an application for passive point-to-point packet loss monitoring, which relies on packet capture in two points in a network. Packetloss runs on top of DiMAPI (Distributed Monitoring Programmable Interface) and is based on its function called EXPIRED_FLOWS, which counts packets captured per flow and returns this information after the flow is expired. A flow is considered expired, when it is not active for 30 seconds (it differs from standard Netflow semantics in that there is no active timeout). Packetloss receives expired flows from remote sensors. When it gets the same flow from multiple remote sensors, it calculates the packets lost.
Features:
- Can monitor real packet loss of user traffic, rather than packet loss of testing traffic, which can be very different in its volume and dynamics
Issues:
- Works only if all traffic between selected IP subnets pass both monitoring points, this should normally be the case when we put monitoring points on the border links of a network, such as NREN - GN2 links.
- Works only when we can capture all packets (with zero loss).
- Currently works only for two-point monitoring (it does not work for full mesh of multiple points).
FORTH has originally developed this application in the LOBSTER project.
Measurements
- User interface for test installation of Packetloss in CESNET is accessible at https://lego.cesnet.cz/packetloss (requires username and password send mail to contact below).
Installation
Packetloss is distributed along with DiMAPI. See [1]. Download the latest version from SVN. Go to applications/packetloss subdirectory. Read INSTALL file there.
Return to Passive monitoring
