PingER phone meeting between HEPNRC & SLAC 8/18/99

Rough notes by Les Cottrell
Attendees: Les Cottrell, Warren Matthews - SLAC; Bill Lidinsky and Shiqui He- HEPNRC.
Comments in bold face italics are action items.

SLAC Home Page

Old Action items

Transparencies for Lepton Phton conf

These were given to Matthias and incorporated into his talk.

HEPNRC staffing progress

Bill has got through the requistion for the new hire in record time with help from Matthias Kasemann. He believes the alternate choice candidate is still interested and hopes to make an offere soon. Bill make an offer to the candidate.

Abstract for CHEP00

There has been a call for papers for CHEP 2000. We briefly discussed presenting one or more papers as we have at the last few CHEPs. The deadline for abstracts is 15 October, 1999. Warren has put together a brief abstract:
How No Collisions Makes Better Physics: All you ever wanted to know 
about achieving your Wide-Area-Networking Goals.

Internet connectivity is critical infrastructure for modern HEP 
experiments, computing goals can only be achieved with reliable, 
fast connections between geographically separate regions.

In this talk we will discuss past and future performance, how the 
requirements of HEP can be met and how we will know. We will examine
several networks and review several monitoring projects.

We aim to detect the recipe for success.
Les also put together a rough outline:
Active End-to-end Internet Performance Measurements in HEP & Elsewhere: 
when no collisions makes better physics

Why we care
        Distance independent collaborations
Methodology, deployment
        thruput, http
Comparison with other Active IEPM projects
        Long term thruput trends
        Comparisons between networks
        Importance of routing
        Monitoring of major HEP collaborations
Bill and Shiqi had not had a chance to review the above and agreed to read and send responses by email in the next few days.


TCP bandwdith metric

Warren is adding a TCP bandwidth metric to the choices. It is based on the Matthis et. al. formula that
BWTCP < 1460 / (RTT*sqrt(loss)),
with a zero loss being set to the lowest measurable based on the sample size.

New beacon sites

Les has issued a new set of beacon sites. It has been installed by 8 of the 19 active PingER monitoring sites. We need to get everyone to install. There was a problem with complaints from with pinging that host in Puerto Rico, so it has been removed from the beacons. HEPNRC stopped pinging yesterday. There was a security incideent which delayed HEPNRC installing the new beacons. After Shiqi makes the modifications it will be a couple of days before it gets reflected in the SAS database. Shiqi will install the new beacons list.

SAS issues


Warren has talked to some SAS people on the technical issues of interfacing to SAS via JDBC. The SAS people were helpful and encouraging. He is now trying to understand the pricing. Shiqi said that the package for Unix costs $18K and can't be separated from the package it comes bundled with.

Warren's student has made encouraging progress with providing JDBC access to some other databases.

HEPNRC are going to look at running the SAS database on Windows NT. This will enable them to use their existing licenses for the Internet access server. It will mean that the archive data will need to be copied to the NT server. Then we will need to test to see that the JDBC access works remotely from SLAC, and if successful we can migrate away from the SAS Unix. Should we need remote logon capability to this machine then we may need to look at the Citrix Windows Terminal Server and running the SAS application under it, however, at this time such access is probably not needed from SLAC. Bill feels this is an October timeframe issue. We should make this a regular agenda item for the next few meetings.

Next Meeting

There will be a phone meeting on Wednesday 8th September at 2:30pm PDT.

Back to Top

Revised 18 August 1999
Comments to