MASSACHUSETTS INSTITUTE OF TECHNOLOGY HAYSTACK OBSERVATORY 18 April 2002

advertisement
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
HAYSTACK OBSERVATORY
WESTFORD, MASSACHUSETTS 01886
Telephone:
Fax:
978-692-4764
781-981-0590
18 April 2002
TO:
Distribution
FROM:
Alan R. Whitney
SUBJECT:
18 April 2002 e-VLBI telecon summary
Attendees:
Lee Foster, Pat Gary, Chuck Kodak, Bill Wildes – GSFC
Tom Lehman, ISI
Steve Bernstein, Lorraine Prior – Lincoln Laboratory
Richard Crowley, Kevin Dudevoir, Hans Hinteregger, Arthur Niell, Alan Whitney – Haystack
Observatory
This telecon is one of an ongoing series of telecons to prepare for gigabit/sec e-VLBI
demonstrations between NASA GSFC and MIT Haystack Observatory using a combination of
network facilities including all or part of Glownet, Bossnet, ISI-E, SuperNet, Max and
GSFC/HECN.
Status Reports
The attached figures of the e-VLBI path have been updated to reflect current status and are pretty
much self-explanatory; critical status items are indicated in red. In addition, the following
comments are relevant:
− Lorraine reported that the Summit 5i at C1 and E3 have been replaced with Alpine
3804’s and a third λ has been added to the D1/D3 link. E11 no longer exists; test
workstation E5 now connects directly to E3. Because one of LL radars is currently down
for maintenance, e-VLBI can temporarily have a dedicated λ for testing (til mid-May
2002), which is much appreciated and should help speed up testing.
− Tom reported some possible problem with one of the Bossnet transponders. Tom and
Steve will consult with Peter Schultz at LL regarding this issue.
− Steve reported that link F1 was severed on 9 April due to a car hitting a power pole in
Somerville, MA. NStar has slightly re-rerouted the link between LL and 230 Congress
St. Steve does not believe this is related to the possible Bossnet problem.
− Kevin reported UDP transfer from Haystack to ISI-E (G7) of ~850-900 Mbps with ~0.5%
drop rate; UDP has not yet been tried from ISI-E to Haystack. Above 900 Mbps, drop
rate rises quickly; at 980 Mbps, drop rate increases to ~5-7% and is quite repeatable.
TCP throughput is often <100 Mbps, though occasionally has climbed to as high as 400
minutes11.doc 4/19/2002
1
Mbps. Kevin has sent Tom a list of symptoms.
− Tom will work with Kevin next week to try to resolve the poor TCP performance
between Haystack and ISI-E. Two different workstations (one dual Pentium and one
dual Alpha) at ISI-E have been involved in the tests, both showing similar performance.
Tom suspects some tweaking is needed.
− Pat reported that H9/H11 link will be in place by 26 April; H9 connection to H7 (Juniper
M160). To make this link, Jerry Sobieski will retrieve an interface (cost ~$25K!)
currently at ISI-E which he has agreed to put into H7 for these e-VLBI tests.
− Pat will get details on test workstation at J1 from Jerry.
− Testing from Haystack to GSFC/Bldg 28 can be conducted through the H8/J4 link
through the Cisco 4912 (J3), but at reduced performance since the 4912 does not support
jumbo frames. Workstation at K6 position can be used for this purpose.
− Pat reported that test workstation at K6 will be moved to GGAO by 26 April. Position at
K6 can probably then be ‘backfilled’ with another (as yet unidentified) G4 Mac test
workstation. Haystack would like to hold onto Bill Wildes Dell test workstation
(currently at B3) for a while longer since Haystack has no other suitable machine at the
moment.
− Alan asked anyone with recommendations for a good PC motherboard for supporting
GigE; must also support 64-bit/66-MHz PCI. CPU in current motherboard (Dell
motherboard with 1 GHz P3) nearly saturates during simple performance testing and
must be upgraded.
e-VLBI workshop at Haystack Observatory
A 2-day international e-VLBI workshop was held at Haystack Observatory on 8-9 April (last
week), attended by ~70 VLBI and network experts. Meeting information and presentations are
available on-line at http://web.haystack.mit.edu/haystack/vlbisystems.html; a meeting summary
will be posted soon at the same address.
Other
Pat suggests that we think ahead to how to keep the link alive after the conclusion of
demonstration period, when loaned and borrowed equipment will inevitably have to be returned.
In particular, the H9 connection into H7 (Juniper M160) and the G11 connection into H3
(Juniper M160) are temporary and need to be solidified with real dollars. Tom suggests we
should get together with NSF and/or DARPA to establish a follow-on program to this
demonstration, which may very well include procuring the equipment to replace currently loaned
or borrowed pieces.
Bill Wildes indicated that Jim Fisher (NASA?) is hosting a round table discussing future
directions of computing for NASA Code Y. A future round table may include high-performance
network discussions, and we should keep our eyes open for this. Bill also indicated that John
LaBreque of NASA HQ is a big supporter of this e-VLBI demonstration experiment and is very
interested in possibly launching a new initiative under Code Y to further e-VLBI development.
minutes11.doc 4/19/2002
2
Pat suggests focusing on at least two things in future activities: 1) solidifying activities at 1 Gbps
and 2) push the envelope to 10 Gbps as next generation. Tom indicated again that he expects to
have 10GigE up between ISI-E and LL sometime this summer.
Pat mentioned that Gary Carter of UMBC (ex-LL) had contacted him, indicating that he (Carter)
is expecting some NASA money to work on high-speed networking problems that NASA may
have. When Pat mentioned the e-VLBI work, Carter was very interested and thought he
(UMBC) might be able to contribute. Steve said that he recalls that Gary is oriented more
toward physical-layer problems. Pat will follow up on this with Gary.
Action Items
Pat: Make connections H9 and H11 by 26 April 2002.
Pat: Move test workstation K6 to GGAO by ~26 April 2002.
All: Think about follow-on activities.
Next telecon
Next telecon will be Monday, 6 May 2002 at 2 pm.
xc:
minutes11.doc 4/19/2002
Steve Bernstein, LL
Jim Calvin, LL
Lorraine Prior, LL
Leslie Weiner, LL
Herbert Durbeck, GSFC
Lee Foster, GSFC
Pat Gary, GSFC
Chuck Kodak, GSFC
Kevin Kranacs, GSFC
Paul Lang, GSFC
Aruna Muppalla, GSFC
Bill Wildes, GSFC
Dan Magorian, UMCP
Tom Lehman, ISI
Jerry Sobieski, Max
Richard Crowley, Haystack
Kevin Dudevoir, Haystack
Hans Hinteregger, Haystack
Arthur Niell, Haystack
Joe Salah, Haystack
3
Westford Antenna
Millstone RPE Bldg (Lorraine)
Dedicated no scheduling
GigE
WS/
Mark 5
Summit
1i
1
A2
A1
GigE
8
Alpine
A4
3804
A3
Dedicated no scheduling
28-1A
14
B1
GigE
GigE
Dell Poweredge 2500
Summit
15
28-1B
B8
B6
B7
Hub
B10
B9
To HS switch
B3
WDM
GigE
5i
100 Mb
B4
C3
16
13
WS/
Mark 5
GigE
B5
B2
Corr
Cntl
Cmptr
GlowNet
Congress St
Boston
(~50 km)
Amp
D2
Lincoln Lab (Lorraine/Steve/Jim)
Schedule to be
coordinated with LL
Cisco
1500
Connection made for e-VLBI
E3
GigE
E2
E1
Alpine
3804
E4 GigE
E5
Test
WS
GigE
E6
(Shared with LL on
scheduled basis)
Cisco
1500
GigE
B11
D3
WDM
LL
C2
Haystack Correlator
WS/
Mark 5
Schedule to be
coordinated with LL
C1
3l (2 GigE + 1 OC-48)
GlowNet
(~100 km)
D1
Bossnet (Steve/Jim)
ISI-E
Scheduling on
Bossnet calendar
E7
Juniper OC-48
1310
M20
E8
GigE
MIT
Campus
WDM
F1
155x
Amp
Amp
F2
E9
Other users (no control)
To be upgraded to OC-48
evlbi011.cdr
Dell Poweredge 500SC
(provided by Haystack)
18 Apr 02
Figure 1: e-VLBI Path - Haystack to ISI-E
ISI-E (Tom)
Scheduling on
Bossnet calendar
MAX @ ISI-E (Jerry)
Need details
Alpha
Test
WS
WDM
Switch
G5
G7
Tuner
Bossnet
Summit
5i
GigE
G6
G8
G6
GigE
H5
H1
G12
GigE
MAX performance node;
need details
Test
WS?
H2
GigE
No scheduling;
sufficient capacity
for e-VLBI
H4
Backup plan
OC-48
Trans-
G2
ponder
G1
OC-48
G9
G11
Juniper
M40
G3
OC-48
Borrowed OC-48
interface - may not
be able to keep
G10
H3
H6
MAX/UMCP (Jerry/Dan)
To be installed by 26 Apr 02
H7
Juniper
M160
GigE
H8
GigE
H9
Primary e-VLBI path
No jumbo frames
J3
Test
WS?
J1
Need details
J2
OC-48
Juniper
M160
Other users ( no control)
Cisco
4912
Switch
GigE
Summit
5i
H10
GigE
J5
J4
WDM
GigE
on loan from GSFC
H11
2 GigE
LuxN
J6
To be installed by 26 Apr 02
LX (~14km)
GGAO/Bldg 201 (Pat/Chuck)
GSFC/Bldg 28 (Pat/Paul/Kevin/Bill F.)
GigE
K2
WDM
Will try to schedule
either K2 or K3
exclusively for
e-VLBI experiment
G4 Mac with
Yellowdog Linux;
need details
(Bill Fink)
K3
K1
K6
Test
WS
GGAO Antenna Trailer (Bill W./Chuck)
K4
Summit
5i
GigE
LX (~7.25km)
K7
1310
Dedicated - no
scheduling necessary
GigE Summit
L1
5i
(~500m)
L2
K5
Test
WS
GigE
L3
SX
L4
L5
K6A
Windows PC
(Lee Foster)
To be moved to GGAO by ~26 Apr 02
Figure 2: e-VLBI Path - ISI-E to GSFC/GGAO
Mark 5
L6
Test
WS
(K6)
evlbi012.cdr
18 Apr 02
Download