Status Report - May 2003

• Activities extending the Network Analysis Infrastructure (NAI) in support of new and developing HPC needs
• Development and distribution of measurement and analysis tools
• Outreach, application support, utilization improvement, and documentation activities
• Ongoing measurement and analysis, networked data, and infrastructure support

Activities extending the Network Analysis Infrastructure (NAI) in support of new and developing HPC needs:

We continue to make progress with OC48, GigE, and OC192/10Gig measurements for PMA and AMP.  

• The nai-p-nca site (National Center for Atmospheric Research, NCAR PMA site) GigE machine is up and should be able to collect traces.  [Bud Hale]
• Setup was done for the AMP monitor planned for the SURFNet site in Amsterdam, the Netherlands (on GigE). [Bud Hale]
• Worked on the newly available MAX OC48MON monitor, which is almost ready to deliver data, but there is too much noise on one of the ports, have been in contact with Dan Magorian of the MAX GigaPOP. [Jörg Micheel, Bud Hale]
• Continued discussions with two Japanese research groups interested in participating in PMA. One is Sony, member in WIDE, the other is APAN.  All the high end monitors are in discussion, OC192MONs and OC48MONs. [Jörg Micheel]
• Continued progress regarding the purchase of additional Dag6 (OC192) cards.  [Ronn Ritke, Jörg Micheel, Mary DiMeglio]

~ New (and developing) strategically important measurement sites ~

• Preparations were finished and new PMA monitors were shipped to the KISTI site (Korean Institute for Science and Technology) and the AMPATH site at Florida International University in Miami. Notification of arrival was received for both machines. Installation to follow at both sites. [Jim Hale, Bud Hale]
• There was activity regarding several new AMP machines this period, many of which will be in the international mesh, others will be deployed in strategic locations.    
• amp-eltn (ELTNet at Eotvos Univ., Budapest, Hungary) was initiated. [Bud Hale]
• AMP monitors for the AMPATH site at Florida International U. in Miami and RNP Network in Rio de Janeiro, Brazil were configured and shipped. [Bud Hale, Jim Hale]
• The site at Great Plains Network GigaPop is preparing for an AMP monitor.  [Bud Hale]
• We have received the confirmation request for an AMP monitor in Mexico.  [Bud Hale]
• Preparations began on an AMP machine for deployment on the Starlight end of the Nauka-Net link to the U.S. [Ronn Ritke, Bud Hale]

IPv6:  Some quite substantial performance increases at the University of Oregon were observed using the IPv6 mesh. I exchanged emails with Bill Owens (NYSERNET) and Joe St. Sauver (Univ. of Oregon) regarding this.  Have not yet found out from Joe St. Sauver exactly what changed, although we can see an IPv6 route change at around 10pm 13 May 2003 PDT.   May 13     May 14     May 19     [Matthew Luckie]

Development and distribution of measurement and analysis tools:

Progress on the reimplementation (repackaging) of AMP and the new testing architecture continued with work on the measurement daemon.  There are currently more than 5000 lines of code for the measurement daemon.  All the address structures were converted to be IPv4/IPv6 compatible.  A test program to check for memory leaks was written and run. Portability issues were examined, tested, and debugged on several OSes. What has been completed has been ported to, and is working on, Linux, FreeBSD, NetBSD, SunOS 5.8 and 5.9 (which are quite different). [Tony McGregor]

Work began on the amplet/amp communications program.  We can now create a certificate authority, issue certificates, and open an encrypted authenticated connection.  Work has begun on the xfer code (that will transfer data between the amplets and the collectors).  [Tony McGregor]

Work continued on developing ipmp_pathchar. An experiment with it in the face of cross traffic was performed.     Documentation and results   [Matthew Luckie]

Further investigation into using nettimer techniques with IPMP to measure and/or profile the L2 capacity of a single hop. [Matthew Luckie]

Measurements to profile the behavior of CRCnet's quickbridge wireless link were taken.    Results    [Matthew Luckie]

Additional functionality was added to Scamper to enable it to continuously traceroute, and to detect when the address list changes.  [Matthew Luckie]

Work towards the development of real-time analysis for the PMA data continued. Preparatory work for creating an algorithm for interpreting the tsh files was done. Improvements were made to the network code including a command line option for users to access host for connection and porting. Issues regarding how to implement a real-time system were considered, (e.g., interpreting data directly from DAG cards vs. using trace files,). [Chris Gross]

Work continued on the reimplementation of Cichlid, including establishing a better queueing system for network traffic and continued designing, building, and coding of the class hierarchy. Designed a ConnectionManager class for the server-side and worked on the DataSet class. The prototyping of a GUI for Cichlid was begun and a mock-up created. I am currently working out how persistent storage of data will work so that users can replay their data streams repeatedly. [Ben Reesman]

Outreach, application support, utilization improvement, and documentation activities:

In Cambridge UK, met with several very interesting people: Jon Crowcroft, Christophe Diot, Carter Bullard, Anja Feldmann, and Nevil Brownlee. Staying with Ian Graham during those days. Discussions surrounded the utility and impact that passive measurements can have and what this group can do to advance the current state of the art. Good talking to Anja Feldmann, we have reached an informal agreement on how to continue our idea on passive data collection at Leibnitz Rechenzentrum in Munich. [Jörg Micheel]

In Baltimore met with Greg Cole, NaukaNet (and NCSA), Steve Weinstein (formerly NSF), and Bill Chang (NSF Program Director for Asia/Pacific) to see the christening of a new cable plowing ship built for Tyco Telecommunications. Greg and I got to talk about his needs for monitoring and how we are going to approach them. [Jörg Micheel]

Problems with the PMA server were first noticed by a number of users who were keen to download more OC48c traces. For those with desperate needs the data was made available in a different location. Have been busy restoring the long traces archive from the HPSS. Although time consuming, it was very refreshing to see how many people really *needed* the trace files, as a matter of urgency. During those 4 days, I had five people asking for OC48c traces and long traces to be made available quickly, as they were working on some paper or contribution with a deadline. I mirrored Auckland-6 and a portion of Abilene-I for folks. Inquiries came from UNC, USC, UMASS, UWISC, from universities in the UK and Brazil. [Jörg Micheel]

Email from Nick Duffield at AT&T Research asking for data collected during the recent Slammer worm attack, January 25th. I was able to retrieve the traces of that day from the HPSS. Really like the utility for this case, I know of a number of other data collection projects who have not been in a position to help Nick in this case. The regular collection by PMA is clearly showing an advantage in postmortem forensics. [Jörg Micheel]

Work continued on the visualization project for AMP to dynamically render sites and serve as a basis for rendering real-time performance of AMP sites.  [Cooper Nelson, Ben Reesman]

Work progressed on the new AMP "front" pages with interactive maps of the AMP sites (U.S. and international versions). [Ben Reesman, Maureen Curran]

Work on the stats analysis of pma.nlanr.net (Web log) continued.  [Cooper Nelson, Jörg Micheel]

Redesigned the MNA Web page layout/style, including a redesign of the navbar (to include alternating images); created templates, and checked for cross browser compatibility. Developed a special AMP related navbar (modeled after new MNA one) for use on the AMP pages.  [Maureen Curran]

Designed the Collaborations and Citings Web pages.  [Maureen Curran]

Small patch to IP to AS address translator in response to user comment.  [Tony McGregor]

Discussion and planning for our activities at SC2003 began.  [Ronn Ritke, Mike Gannis]

Ongoing measurement and analysis, networked data, and infrastructure support:

An extended abstract (6 pages) was submitted to the SIGCOMM Internet Measurement Conference (IMC) 2003, entitled "On Segmentation of Internet Paths for Bandwidth Estimation". [Matthew Luckie; Tony McGregor (ed.)]

Leipzig-I, a new data set, has been collected by Klaus Mochalski at the University of Leipzig. Next is to make it ready for publishing on PMA. [Jörg Micheel]

Investigation into the possibility of using "solid disks" for the AMPlets continued, leading toward the use of compact flash memory technology. Cost comparisons and efficacy testing to follow. [Bud Hale]

Work continued regarding the creation of the third AMP server, OHM. Am_slave is now working on OHM. The data collected from amp-sdsc mirroring on AMP and VOLT is also mirroring to OHM. OHM is ready for more extensive testing. [Jim Hale]

After a number of PMA monitors stopped collecting data, it was determined that one of the main culprits for the hangs are the temperature sensors, which appear to be nothing but a nuisance, so were disabled.  [Jörg Micheel]

Due to the problems with the PMA server, planning discussions and investigations were begun regarding replacing/upgrading it. The server is nearly five years old; the last downtime for the machine was December 2001.  [Jörg Micheel, Jim Hale, Bud Hale]

Trace collection began again the Rice U. site (naitxs).  [Bud Hale, Jörg Micheel]

Took care of arrangements to bring Klaus Mochalski to SD for the summer (he will conduct PMA research while here) and send Chris Gross to New Zealand to work on PMA research with Jörg while there.  [Ronn Ritke]