Participation in the Network Analysis Infrastructure (NAI) as a remote site: What Happens?

Researcher:   Bud Hale, Systems Administrator, NLANR Measurement and Network Analysis Group

What happens when a researcher, or the information systems manager at an institution supporting research, decides to participate in the NLANR Measurement and Analysis Group's activities by hosting a network monitor? This article briefly explains some of the benefits and the procedures involved.

The NLANR Measurement and Network Analysis Group has created, and continues to expand, a network analysis infrastructure (NAI) to support network research and engineering of high performance research networks. The NAI includes a passive monitoring project (Passive Measurement and Analysis [PMA]), an active monitoring project (Active Measurement Project [AMP]), and the corresponding network management and control data necessary to maintain this infrastructure. Together these projects have deployed nearly 140 monitors (predominantly active monitors, currently numbering 116) around the high performance community (HPC) in the United States.

To examine the data collected and gain an understanding of how institutions contribute to the research and how the data can be used, as well as the benefits to institutions (and systems managers) of having access to performance data regarding their network, please see our Web site. Resources are listed on this page. Web pointers to our active and passive measurement research are listed on the left; additional information about participating as a monitor host site is available on the right.

Requesting and receiving a monitor:   procedures and specifications

Among the first steps is to identify the networks connected to the institution's research programs and departments. This self examination should result in information upon which decisions can be made regarding physical location as well as IP address assignment for host IP address, network address, subnet mask and default gateway IP address. All of this information is necessary to complete the Web-based request form.

Related to physical location considerations are factors involving chassis size, rack space and ethernet network connection cabling. Here one needs to decide on the correct network port location as well as appropriate cabling to that port. The monitor chassis is a 19 inch rack mountable chassis requiring seven inches of vertical rack space in a cabinet allowing for an 18 inch deep chassis plus room for rear cable connectors. A minimum of approximately 20 inch cabinet depth is recommended. Our Web pages provide pictures of the monitor chassis and the hardware specifications (see thumbnail image and pointers to the right).

Armed with the results of the decisions discussed above, the candidate participant is now ready to submit the participation request for the desired monitor (active or passive). A Web-based request form/ questionnaire is available (see link in sidebar to the right); please use the comment fields to describe connections and any special needs and/or expectations.

Upon receiving the request, the information submitted enters the NLANR Site Information and Status database. This is followed by the preparation of a monitor machine for shipment. Preparation of the machine includes editing the BIOS setup to insure that when installed, the machine will reboot following loss of power. A standard FreeBSD system disk is installed and the network startup files are edited using the submitted site host information. A sysadmin account is installed such that site systems managers may exercise control if needed. As shipped, the machine is fully prepared and operational, needing nothing more than mounting and connection to power and the assigned network port. After the monitor arrives at the remote site, is unpacked, mounted, connected and powered, the site personnel need to notify the NLANR program systems administrator, Bud Hale. This can be done by phone or email message (contact information is in sidebar to the right). If the connections are good, no other action by site personnel is necessary.

After testing for reachability by the NLANR systems administrator located at the San Diego Supercomputer Center (SDSC), certain site specific software is loaded to the machine and started. This is followed by addition of the site host information to the servers in the data collection systems at SDSC and the addition of the site to the NLANR site map. Thus, the new participant's institution is added to the infrastructure and site network data and statistics begin and become available for review.

back to top

To subscribe, or if you have comments and/or questions, please contact us.
© 2000    The NLANR Measurement and Network Analysis Group,  located at the San Diego Supercomputer Center (SDSC), University of California, San Diego (UCSD).   This work is supported by the National Science Foundation (NSF) (cooperative agreement no. ANI-9807479). Any opinions, findings and conclusions or recommendations expressed in this publication are those of the author(s), and do not necessarily reflect the views of the NSF.