1/ What is the IEPM ?

   The Internet End-to-end Performance Monitoring (IEPM) Group has its
   origins in 1995 in the WAN Monitoring group at SLAC, which monitored
   links to many sites that SLAC collaborates with by using the IP Ping
   facility. The ESnet Network Monitoring Task Force (NMTF) quickly
   become involved in the project and the work was extended to cover ESnet
   sites. Several DOE Labs installed the PingER monitoring software which
   greatly improved our knowledge of the network. In April 1997 the
   International Committee on Future Accelerators (ICFA) created the
   Network Task Force to cover the needs of the High Energy and Nuclear
   Physics (HENP) community. In 1998, the Cross Industry Working Team
   (XIWT) adopted the PingER tools developed out of the above efforts and
   used them to monitor their member sites. 

2/ How does it work ?

   We use the standard ICMP ping facility to provide a measure of
   end-to-end link performance. Ping is a simple tool which comes
   installed on most platforms, and hence there is no complicated
   software to be installed. We use ping to measure round-trip performance
   including response time, packet loss, reachability, and  
   unpredictability.  Every 30 minutes, we ping a set of remote nodes with
   11 pings of 100 data bytes each. The pings are separated by at least
   one second, to help reduce self correlation effects between individual
   pings, and the default ping timeout of 20 seconds is used. The first
   ping is thrown away to reduce possible effects such as priming of
   caches. The minimum/average/maximum for each set of 10 pings is
   recorded. This is repeated for ten pings of 1000 data bytes.

3/ Does this give a good indication of network performance for
   "real" applications ?

   The easiest way to validate ping is to demonstrate that measurements 
   made with it correlate with application response. Such a correlation
   is clearly seen when comparing HTTP GETs and ping responses.
   
4/ How do I join the effort ?

   Send email to iepm-l@slac.stanford.edu for details.

5/ Are the tools and software available ?

   Yes, visit http://www-iepm.slac.stanford.edu

6/ What makes 'pinger' preferable in this work to Caida tools like 'skitter' 
   and 'skping'?

   The two measurements are complementary. The goals of skitter and 
   PingER are different. Skitter is trying to characterize the Internet, 
   whereas PingER is trying to characterize end-to-end performance 
   between selected sites of interest to some community. 

   As a result skitter makes path measurements between a larger 
   number of pairs of sites (similar number of monitoring sites 
   (PingER about 30, Skitter about 20) but Skitter has many more 
   remote sites, 60,000 eventually). Skitter makes the measurements
   on roughly an hourly basis whereas PingER is on a half hourly basis.
   Skitter uses a single packet size (52 Bytes) whereas PingER uses 100
   and 1000 Byte packets. Skitter gathers RTT measurements and
   also information on individual nodes along the path (like traceroute).
   PingER only makes measurements to the end node. 
   PingER is more interested in time series (in particular long term) 
   for the given pairs and has data going back to Jan 1995. In many  
   senses time is nonsensical for skitter measurements in most cases 
   since it is not possible to obtain a path measurement instantaneously 
   without simultaneous access to all IP devices along the path, 
   skitter tends to provide snapshots.
 
   Skping on the other hand appears to be used as a high resolution tool 
   (i.e. frequent measurements between 2 hosts) to understand/trouble shoot 
   end-to-end performance. I do not see any long term archive of measurements 
   that can be accessed for further analysis. PingER has a publicly 
   accesible long term archive and is lower resolution (and lower impact 
   on the remote sites and so can be used to get information on sites that 
   have poor connectivity, e.g. in developing nations.

   That said skping does have some very nice canned visualization tools and 
   has paid a lot of attention to the timing. This means that the 
   skping measurement engines are probably uniform (or at least all running 
   the same OS with the appropriate modifications) and possibly are 
   centrally managed. PingER lets the monitoring sites choose their own 
   measurement engine, which is often shared with other tasks and the 
   monitors are usually managed by the monitor sites. Also PingER, since 
   it makes the data selectable and downloadable via the web in a format 
   suitable for many analysis packages (such as Splus and Excel), it allows 
   interested  users to make their own analyses fairly easily.


   You have to be a paying Caida member to have access to skitter and skping,
   whereas access to the PingER tools and data is free.