Figure 1. RRDtool-based visualization of July 2001 PMA data. more detailed view
The development and ongoing management of the network analysis infrastructure (NAI) continued with excellent progress again this quarter.
Device obsolescence and technology advances continue to be factors for both the Active Measurement Project (AMP) and the Passive Measurement and Analysis (PMA) project. Changes in configuration and in site locations have been necessary in order to keep pace with network growth, rapidly changing networking technology, and the resultant changes in the high performance network infrastructure.
Major work was done to the PMA trace data collection and storage infrastructure. After discussions and planning, a decision was made to introduce the trace anonymization on the monitors and that all of the raw trace files would be removed from the NAI server. Also, the major project reimplementing many of the PMA tools from Perl to C continued successfully with another two analysis scripts fully reimplemented.
As in previous quarters, the AMP data collection functionality continues with a very high level of reliability and minimal outages. This demonstrates the maturity of the infrastructure, support systems, and procedures. Additional sites have been added to the AMP infrastructure; while few in number, they are in important locations.
Documentation and tool development for all aspects of the NAI continues to progress and develop with a number of new tools and Web pages created this quarter for all of our activities. These help to improve the ease of use and accessibility of our data, analyses, and tools for other researchers, network engineers, and students.
Significant work was done to the PMA trace data collection and storage infrastructure. The new data server will also provide more graphical information on the amount and types of data which are collected daily.
Two papers from our group were accepted for presentation at the ACM SIGCOMM IMW2001, Internet Measurement Workshop. Greater access to the published papers of the group was achieved through a revamp of the publications list which included the addition of links to authors' versions for each of the papers.
Attendance at workshops and meetings helped begin new (and strengthen ongoing) collaborations. The PMA staff is especially interested in enriching the available trace data via collaborations. We hosted the High Performance International Internet Services (HPIIS) Workshop, in August; participants came from all over the world wide high performance community.
Students continue to make significant contributions to the project in many areas (script writing, displaying data graphically, Web page development and maintenance, technical support regarding data, and other activities).
For PMA activities, a primary focus this quarter was to enrich the set of available trace data via collaborations. New traces were expected to arrive from the Bell Labs Internet access link as well as a first multipoint trace from the TransPAC. For various reasons installation of the measurement points failed. Instead, PMA has managed to get access to a first local multipoint trace (at the University of Auckland).
Moderate progress has been made on the router instrumentation. The Waikato Applied Network Dynamics (WAND) research group has been able to demonstrate OC48c network capturing at a commercial Internet backbone link in San Jose. The set of eight Dag4.2 cards is expected to become available sometime in November 2001. We are preparing and testing a suitable computer system capable of hosting a pair of these cards in a 2U rack-mounted chassis, (size is a factor due to the anticipated space constraints at the Indiana GigaPop Abilene router [IPLS]). A new monitor machine is being created for use in the Dag4 card support development. The plan is to ship it to Waikato (before the end of October).
A new site was deployed in August (at the University of Buffalo). We decided to retire two OC12c monitors at the vBNS - 12NCS and 12SDC, which hardly saw any traffic in the last year or so due to discontinuing use of the vBNS at SDSC and NCSA. Also retired was another system at the SDSC FDDI ring and the Cerfnet access link. We are planning to introduce new monitors at SDSC (OC12 connections): one to the Abilene network and the other to the CalRen network. These connections are expected to yield very useful packet traces. As a stepping-stone we are looking into implementing a small scale router instrumentation at SDSC with a total of four links at OC3c and OC12c. This will require some type of time synchronization; preparations are under way to bring a GPS signal into the machine room for this purpose.
Passive monitors now being created use the TYAN 2510 system board, which is expected to be used in the monitors deployed with the remainder of the Dag3 optical monitor cards. Approximately 14 Dag boards remain to be delivered by Waikato; these are expected to be the newer Dag3.5 cards. The Dag3.5 design will have a wider dynamic input power range of -14 dbm to -29 dbm (vs. the -26 dbm of the current Dag3 cards). The 3 dbm lower signal range should do much to ease the optical connection difficulties experienced with the current Dag cards.
As a result of retiring some AppTel POINT cards we have found a new collaborator at the National Library of Medicine, and are expecting some longer trace files with the loan of a pair of cards to Victor Cid.
Major work was done to the PMA trace data collection and storage infrastructure. Discussions during Joerg Micheel's visit to SDSC in August resulted in the decision to introduce the trace anonymization on the monitors and that all of the raw trace files would be removed from the NAI server, (which has been turned into pma.nlanr.net). As a result of this, we have gained twice as much disk space, allowing us to make both the daily traces and the long trace data sets available from the pma server. In addition, the new server provides consistent access via http and ftp protocols. The reconfiguration, resulting changes to the Web interface, and changes to the monitor data collection occupied all of September and is an ongoing process. As of the end of September, a total of six machines are capturing data on a daily basis, with more monitors being made available every day.
Progress has been made in rewriting the PMA postprocessing tools and versions of those are expected to be integrated into the data collection during the revamp of the monitor infrastructure. See also the Documentation, Tools, and Publications section later in this report for additional details.
The PMA staff continues to make important contacts for collaboration on the analysis of measurement data and the installation of new sites, and the collection of new traces (including long ones).
Further information on PMA activities is detailed in the documentation/tools, papers, and collaborations sections, following. Please see the Web pages for more information on the PMA project: http://moat.nlanr.net/PMA/.
As in previous quarters, the AMP data collection functionality continues with a very high level of reliability and minimal outages. This demonstrates the maturity of the infrastructure, support systems, and procedures. Additional sites have been added to the AMP infrastructure; while few in number, they are significant additions. For example, the University of California, Santa Barbara (UCSB) has just begun participation. This action was prompted by UCSB gaining a network research group led by Martin Swany. Other sites beginning participation in the AMP mesh this period include the Rensselaer Polytechnic Institute and the Rochester Institute of Technology. Both of these sites involve network measurement collaborations with Bill Owens of NYSERnet.
During the previous quarter, we reported that the system board used in the AMP monitors was becoming obsolete and therefore, unavailable. Tests and investigation into replacement boards resulted in the selection of the ASUS CUSI-M board with the SiS 630E chipset. A side benefit of this change is that the new system board selection allows for the AMP monitor chassis size to be reduced from a 4 rack unit (RU) size (7 inches high) to a 1 RU size (1 3/4 inch). The advantage of needing much less rack space is that it makes remote site installation easier and participation more attractive to new sites.
To improve the performance and stability of the AMP data collectors (amp and volt), the system boards on both were replaced. The new system boards give a much improved performance. Also, the new system boards enabled an increase in RAM on each system to 1 Gigabyte. This allowed a much larger disk cache and consequently improved the disk usage and does much to improve system stability and reliability by greatly reducing disk activity. This change was transparent to data collection and end-users (researchers).
However, there were disk failures this quarter. The redundancy provided by the amp and volt data collectors allowed full recovery with no loss of data and only minimal disruptions of service, but required quite a bit of effort on the part of the AMP staff.
The AMP staff continues to pursue issues regarding AMP data storage space. In more recent months this has involved the utilization of the High Performance Storage System (HPSS) at SDSC. Although this is an excellent solution to the problem of continued growth of data storage needs, some issues remain. In part because of the large number of small files created by AMP, we are only slowly managing to automate this process. Solutions for these issues are being pursued with the HPSS staff at SDSC and resolutions are forthcoming.
Work has continued on IPMP (the IP Measurement protocol proposed by the AMP team - please see http://moat.nlanr.net/AMP/AMP/IPMP/). Progress this quarter included a new implementation of IPMP in the Linux kernel. The application level components of IPMP are expected to operate correctly under Linux, but this was still to be tested at the end of the quarter.
An investigation into timing issues in measurement systems began with an active/passive validation of AMP. This work is still in progress and should be completed during next quarter.
Some analysis of the "Code Red" and the "Nimda" worms and their effect on the AMP sites was performed. A Web page with the results was posted.
Daily archives for the RTT data over a month were created and ftped to Yan Chen at Berkeley for their study into adapting Internet applications based on network measurements (http://www.cs.berkeley.edu/~yanchen/research/wnmms/).
A cgi script was written to extract hop count frequencies from the AMP data for Piet Van Mieghem, Delft University of Technology (TU Delft - http://itu.tudelft.nl) to support their study into the way hop counts change over time.
AMP project leader Tony McGregor has been invited to give the keynote address at SAINT-2002, the 2002 International Symposium on Applications and the Internet, 28 Jan - 1 Feb 2002, Nara City, Nara, Japan
Further information on AMP activities is detailed in the documentation/tools, papers, and collaborations sections, following. Please see the Web pages for more information on the AMP project: http://moat.nlanr.net/AMP/.
A number of refinements were made this quarter to the developing High Performance Wireless Research and Education Network (HPWREN) measurement and analysis infrastructure. Additional Perl scripts were written for use in analysis of existing data. New scripts were also written for stat.hpwren.ucsd, the Web interface which allows users and other researchers to view graphs of throughput on the wireless network (http://calorie.nlanr.net/Wireless/throughput_frontpage.cgi).
The passive hub in the machine room was upgraded to a switch. Collisions had been overloading the connection. The new switch gives a full-duplex connection and no further collisions. Enhancements and modifications were made to the programming methods for the weather station data. The SNMP Tsunami and Lucent pages also have additional scripts and data analysis.
There have been improvements to the interfaces of several existing Web data pages, making them easier to use and more efficient, foremost among them is the router page: http://stat.hpwren.ucsd.edu/HPWREN/New/RouterTraffic/graph_router.cgi.
There were several activities involving the seismic collaboration with geophysicist Frank Vernon (Scripps Institution of Oceanography [SIO])including the alignment of sensors (using dataloggers), GPS activities, and other related tasks. Work began on an "mquake" notification of seismic data via multicast.
Documentation and tool development for all aspects of the Network Analysis Infrastructure (NAI) continues to progress and develop with a number of new tools and pages.
Figure 1. RRDtool-based visualization of July 2001 PMA data. more detailed view
Significant work was done to the PMA trace data collection and storage infrastructure. The new data server will also provide more graphical information on the amount and types of data which are collected daily. To this end, we have begun work on a RRDtool-based visualization, which will see improvements as it is further developed. See Figure 1 above, and http://pma.nlanr.net/PMA/collection/. Plans are underway to also display all of the data collected since the beginning of the project in 1998.
We are developing an illustrated trace archive. Rather than simply downloading trace files then starting the analysis, researchers should already have a good idea of what to expect, and therefore be able to choose trace files specific to their needs. Our collaboration with Bell Labs is the first step of drawing other researchers' analysis results to the data, slowly building up a "knowledge base" of Internet header traces. The hope is that we draw more people into this activity over time.
The major project reimplementing many of the PMA tools from Perl to C continues. Rewrites of the tsh2asc and tsh2bpps analysis scripts were completed, and significant progress was made on crl2tsh, ta2sum, and tsh2plen. (This revamp of the PMA data collection procedures became crucial last quarter due to the outage of our main trace collection server.)
The Site Traffic Summary is a new tool available for understanding the network usage of active measurement data collectors and analysis machines. A select set of servers that are part of the AMP project are included. One server is used for data analysis of HPWREN data and the rest are used for data collection and analysis of AMP data. http://calorie.nlanr.net/AMP/show_traffic.cgi
A new AMP `day in the life' dataset was built; replacing the existing older one. This was done in response to user requests. http://watt.nlanr.net//amp-data/101.7.25.tar.bz2
The IPAS software was completed and made available on the Web site. IPAS is a small package to translate IP addresses into Autonomous System numbers. It is provided along with an NLANR maintained server to make translation from IP addresses to AS numbers easier for the network measurement community. Our database is updated weekly using data from the Route Views Project. http://moat.nlanr.net/Software/IPAS/
A new Cichlid image of the logical high performance connection (HPC) network across the U.S., including Abilene and Internet2, has been developed. See Figure 2 in section 4, following.
The Web page list of papers published by NLANR Measurement and Network Analysis Group members was redesigned and updated. Each of the papers now has a link to the authors' version of the paper (both PDF and PS-GZ formats, if available); all papers from 1998 to present are included on this list (http://moat.nlanr.net/Papers/). Earlier papers are in a directory of their own, linked off the Publications and Resources page (http://moat.nlanr.net/PubsResources/).
A new Web page was created, "Work in Progress," for use in making new and preliminary work available to other researchers. The Publications and Resources Web page and subdirectories were restructured and also redesigned. http://moat.nlanr.net/PubsResources/
Groundwork began on the next issue of the Network Analysis Times, to be published next quarter. The Network Analysis Times was distributed at the HPIIS workshop (SDSC, August), SD-Network Access Point (NAP) Member Meeting (SDSC, July), 27-31 ACM SIGCOMM 2001 (UCSD, August), meeting of the Hispanic Engineers Society (SDSC, September).
Preparations for our presence and demos at SC2001 began. The Abilene Cichlid server was readied for presentation. Additional (old) Cichlid servers were also prepared for SC2001.
Of the six papers (representing nine different members of the group) submitted to the ACM SIGCOMM IMW2001, Internet Measurement Workshop (November 2001, San Francisco), two were accepted for presentation and publication in the proceedings. This was excellent as less than one-third of the more than 100 submissions were accepted.
Luckie, M.J. and A.J. McGregor, and H-W. Braun. Towards Improving Packet Probing Techniques. Accepted by ACM SIGCOMM Internet Measurement Workshop, San Francisco, California, Nov. 2001.
Micheel, J., S. Donnelly, and I. Graham. Precision Timestamping for Network Packets. Accepted by the ACM SIGCOMM Internet Measurement Workshop, San Francisco, California, Nov. 2001.
Matthew Luckie and Tony McGregor gave a talk on IPMP to the Waikato University Computer Science Department.
McGregor A.J. Quality in measurement: beyond the deployment barrier. Accepted for publication, Keynote address, SAINT-2002, the 2002 International Symposium on Applications and the Internet, 28 Jan - 1 Feb 2002, Nara City, Nara, Japan.
The NLANR Measurement and Network Analysis Group was represented this quarter at the following meetings:
JET meeting, August 21,2001
CISCO Workshop, Miami, Florida, August 15, 2001
SIGCOMM Traffic Measurements for IP Operations
ACM SIGCOMM meeting, UCSD, August 27-31, 2001
AMPATH meeting, Miami, Florida, August 16-17, 2001
High Performance International Internet Services (HPIIS) Workshop, SDSC, August 23-24, 2001
We hosted this workshop in August in conjunction with NSF's High Performance International Internet Services (HPIIS) Project. High performance network researchers and administrators from all around the world attended. The workshop focused on the development of measurement metrics to quantify the use of HPIIS links in addition to possibly identifying research collaborations. Domestic and international advanced networks were involved. On the second day, Todd Hansen presented a demonstration on creating your own AMP network. For more information on this workshop, please see the Web page: http://moat.nlanr.net/Workshops/HPIIS-2001/.
A primary focus for PMA is to enrich the set of available trace data via collaborations. New traces are expected to arrive from the Bell Labs Internet access link as well as a first multipoint trace from the TransPAC. In addition, we are developing an illustrated trace archive. Our collaboration with Bell Labs is the first step of drawing other researchers' analysis results to the data, slowly building up a "knowledge base" of Internet header traces. The hope is that we draw more people into this activity over time.
We have numerous continuing collaborations; the following list includes both long-term relationships, as well as recently developed ones. This quarter we continued work with the following organizations and researchers:
Attendance at workshops and meetings helped begin new (and strengthen ongoing) collaborations. We hosted the High Performance International Internet Services (HPIIS) Workshop, in August, at SDSC. High performance network researchers and administrators from all around the world attended this workshop which provided new opportunities for networking and collaboration. We have since confirmed a passive monitor installation at the AMPATH access link at Florida International University. (For more information on the workshop, please see the meetings subsection in the previous "Papers, Presentations, and Conference/Meeting Participation" section, and/or the Web page: http://moat.nlanr.net/Workshops/HPIIS-2001/.)
Attendance at ACM SIGCOMM at UCSD in August proved a very fruitful place to make new contacts and learn about the research work of other groups.
Attendance at PAM2001 (last April) led to a collaboration with Piet Van Mieghem, Delft University of Technology (TU Delft - http://itu.tudelft.nl). A cgi script was written to extract hop count frequencies from the AMP data for him.
Several attempts have been made to find collaborators for collecting data in commercial networks, but these efforts have failed so far. We continue to pursue opportunities, however at the same time, we need to consider making the best use of our existing installed monitoring basis.
Figure 2. Cichlid 3-D visualization of the logical high performance connection (HPC) network across the U.S., including Abilene and Internet2. For additional information, please see section 3. more detailed view
Students continue to make significant contributions to the project in many areas (script writing, displaying data graphically, Web page development and maintenance, technical support regarding data, note takers for the HPIIS workshop, and other activities).
David Cheney (UCSD undergraduate) continues to work in a technical support capacity for the PMA project. This quarter he was primarily involved with the reimplementation of the PMA tools. (Please see Sections 2. "Measurement Infrastructure Development and Management and Network Analysis Activities" and 3. "Networked Data, Documentation, Retrieval and Analysis Tools, and Result Presentation" above for additional detail.)
Stephen Donnelly, a University of Waikato Ph.D. student (Computer Science), is looking into timing issues in measurement systems. He has started doing an active/passive validation of AMP.
Building on Cichlid developer Jeff Brown's initial work, Justin Fields created a new Cichlid image of the logical high performance connection (HPC) network across the U.S., including Abilene and Internet2. (See Figure 2 above.) He will be attending SC2001 next quarter to do demonstrations of this and other Cichlid servers. He maintains the Web page of the Cichlid servers: http://moat.nlanr.net/~jfields/cichlid-servers.html.
Kuo-Wen Lo (UCSD undergraduate) presented his work at the SDSC student poster session this quarter. He displayed a variety of graphs and charts that he developed during the year. He works with the measurement and analysis data from the wireless network (HPWREN).
Matthew Luckie, a University of Waikato Ph.D. student (Computer Science) returned to New Zealand after having spent nearly six months with us at SDSC. While here, he spent time on the AMP project, especially on developing IPMP; he continues to work with us.
Jose Otero began graduate school (PH.D. student in geophysics) at the Scripps Institute of Oceanography (UCSD) this Fall. Jose continues to work with Frank Vernon (SIO geophysicist, HPWREN Co-PI) on seismic activities, including dataloggers, GPS activities, and other tasks.
Neil Cotofana (UCSD undergraduate) who has been working primarily on Network Routing activities with us for over two years, has decided that for his last quarter of undergraduate education, he is going to work with SDSC's Data Mining Group - SKIDL. He previously created a collaboration with SKIDL to explore the application of data mining techniques and commercial tools to networking data. Specifically, GateD BGP routing data (collected and archived by NLANR) was analyzed.
-30-
