-- Main.mbarroso - 20 Mar 2007

This page contains a collection of the answers given by the ROCs/sites to the network monitoring survey as requested by Mathieu Goutelle (SA2).

Asia Pacific

Our current network monitoring status and plans are as follows:

Smokeping: We use smokeping to monitor the network latency, packet loss and jitter to each of the site in the region from ASGC. This allows us to monitor the network performance between the ROC and RC. This can also give us a general idea of the network quality for the RC, but may not be entirely accurate.

MRTG and Weathermap: We use this to collect network usage information only from ASGC managed routers in our international networks.

Nagios: Nagios is used to fault detection of ASGC owned network equipment and APROC RC center connectivity status. If a network fault occurs at any RC, we are notified through email.

In addition we are testing additional tools for network monitoring such as Perfsonar and IEPM. However this is still work in progress.

Please let us know if you need further information for NPM.

Central Europe

Regional installation of NAGIOS is used to detect problems with networking links.

Monitoring of network traffic is done separately by sites. The most popular is Ganglia - almost all sites declare to have it installed. Also, Cacti is popular (http://www.cacti.net/).

Additionally, following tools are used on sites:

• MRTG http://oss.oetiker.ch/mrtg/
• RRDTool http://oss.oetiker.ch/rrdtool/
• NFSen to detect abnormal traffic patterns http://nfsen.sourceforge.net/
• Cricket that measures each port and generic parameters of network devices http://cricket.sourceforge.net/
• ipac-ng used for IP accounting on SE/DPM pool servers for accounting of a dedicated connection from Tier2 to Tier1 site. http://sourceforge.net/projects/ipac-ng/

Some small sites declare to have no network monitoring at all.

CERN

All information is contained in CERN monitoring portal: http://cern.ch/monitoring

DECH

Network monitoring tools used at Gridka: Cacti and Netflow, IEPM

Cacti for collecting port statistics of throughput and error status.

With the help of Netflow we collect specific transfer statistics between endsystems.

In addition, an IEPM monitor is implemented at Gridka (see http://lhc-opn-mon-fzk.gridka.de/iepm-bw.fzk.de/slac_wan_bw_tests.html).

All public monitoring information for Gridka can be accessed here: http://www.gridka.de/monitoring/main.html

Please let us know if you need any more details.

France

Tier1 + Tiers2:

Network usage information: Cricket and Weathermap http://netstat.in2p3.fr

Fault detection and alarm system (java client, sms, email, ...): home made tool http://netsurv.in2p3.fr

Security and performance tools: home made tool http://lpsc.in2p3.fr/extra/ (cf NetFlow)

Tier1:

Other tools: RIPE TTM, IEPM, PerfSonar, EGEE NPM

Italy

Monitoring and alarm tool used at INFN (in particular at tier1/tier2 sites):

MRTG to collect network usage information;

NAGIOS is used to fault detection and alarm system (via email and sms).

NetFlow/sFlow to get top speaker, top network flow, top network application, etc.

GridICE is used to monitor the grid information.

INTN-T1 and INFN-BARI are testing LEMON (only for monitoring purpose) in particulare for their storage system.

At CNAF we are also testing Argus (NIDS).

Northern Europe

No input yet.

Russia

We use the monitoring system developed at the IKI RAS that collects netflow statistics from sites and presents it to the network management staff for analysis.

We use Smokeping to detect the connectivity status to the external networks.

Sites themselves and their NOC units are using the variety of tools to monitor the network links and equipment status: MRTG, Nagios, Cacti, Ganglia and some home-grown tools.

South East Europe

Our current network monitoring status and plans are as follows:

Smokeping: We use smokeping to monitor the network latency, packet loss and jitter to each of the site in the region. This allows us to monitor the network performance between the ROC and RC. This can also give us a general idea of the network quality for the RC, but may not be entirely accurate. http://mon.egee-see.org/cgi-bin/smokeping.cgi

MRTG and Weathermap: We use this to collect network usage information from GRNET and SEREEN managed routers in our international networks.

Nagios: Nagios is used to fault detection of RC center connectivity status in Bulgaria. If a network fault occurs at any RC, we are notified through email and or SMS. If needed we can expand this to cover the SEE Region.

South West Europe

No input yet.

UKI

An overview of network monitoring in the UK is shown at http://www.gridpp.ac.uk/gridpp18/P20070320-RT-18Collab.pdf