RE: data tranfer to USA

[Padiyath Sreekumaran <Kumar.Padiyath@psi.ch>]

  Hi Peter,
  Thanks for your mail. I asked the person incharge of communication.
  According to him we are are not atpresent charging any customer.
  But we would like to know the network load when you retrieve the datas
  (whether this retrieval will cause slowing down the network to our Backup
  of systems)My suggestion is you retrieve few files. Please let me know 
  when you start.Then we can observe it here. 

   Regards,
   Kumar

-----Original Message-----
From: Peter Kammel [mailto:kammel@npl.uiuc.edu] 
Sent: Freitag, 24. Juni 2005 15:39
To: Padiyath Sreekumaran
Cc: Markushin Valeri; Claude Petitjean
Subject: RE: data tranfer to USA


Dear Kumar,

Thanks for your reply. 100MB/s would be fine. I read on
http://ait.web.psi.ch/services/costs.html
that the off peak charges for transfer to the USA 
is CHF 5/GB. If this is real money, I would not transfer
our 1 TB over the PSI network. Could you check with
your network expert, if these are real charges?

Many thanks

Peter




On Fri, 24 Jun 2005, Padiyath Sreekumaran wrote:

> 
>     Hi Peter,
>     Thanks for your mail.The best way is always is to copy the data to
disk
>     (you need 5-6 disks each more than 200GB capacity) and transfer it.
>     Tapes are also good. But you have to copy to tape first and read 
>     from the tape which is more time consuming. We have an external 
> LTO2 tape
>     which we use for our purpose. So we cannot lend it. I asked the 
> network person
>     and he says that at present you can have 100Mbt/sec from PSI. From 
> CERN it may be
>     possible to have 10Gbit/sec. But I donot know whether that helps 
> much since
>     we donot have that transmission rate. I hope this helps you to 
> certain extend. If
>     you need my help in some way please let me know.
> 
>     Regards,
>     Kumar
> 
> 
> -----Original Message-----
> From: Peter Kammel [mailto:kammel@npl.uiuc.edu]
> Sent: Donnerstag, 23. Juni 2005 23:54
> To: Valeri Markushin; Padiyath Sreekumaran
> Cc: Claude Petitjean; Fred Gray; Francoise Mulhauser; David Hertzog
> Subject: data tranfer to USA
> 
> 
> Dear Valeri, dear Kumar,
> 
> Let me write to my old friends concerning this question. If you
> cannot help directly, perhaps you can refer me to the right person at PSI.
> 
> I have successfully applied for an initial CPU allocation at the NSF 
> funded NCSA Xeon Cluster at UIUC ( 1450 processors) 
> http://www.ncsa.uiuc.edu/ for analyzing  the data from our MuCap 
> experiment at PSI. This will provide critical CPU power in addition to 
> the Merlin cluster and our research group's farm at UIUC. An overview 
> of our projected needs are given in my attached email to NCSA. Of 
> course, we are very interested in continuing the use of Merlin and 
> would like to stay informed about its upgrade plans. We will discuss 
> all these matter with you in fall, once we arrive at PSI for an
> 
> extended run.
> 
> A critical issue will be the transfer of the data from PSI to NCSA. 
> NCSA would prefer to have the data sent via high speed backbone 
> connections.
> 
> Looking at the PSI website, even the transfer of a few Terabytes would 
> be terrible expensive. Are there additional High Energy Networks which 
> we could potentially use for this purpose at a nominal charge? Or will 
> those exist in 6 months? Or does PSI have contacts to CERN, which 
> could help?
> 
> Currently, we would urgently need to tranfer 1 TB to the States, but 
> were unable to locate LTO drives at PSI, so as to prepare tape copies 
> for shipment. Is there a possibility to transfer the TB over the net 
> as a pilot project to test this possibility?
> 
> Thanks for all information and best regards
> 
> Peter
> 
> -----------------------------------------------
> Dear Radha,
> 
> We would like to submit requests for Tungsten cluster for processing 
> the data on our from our precision muon experiments: MuCap 
> http://www.npl.uiuc.edu/exp/mucapture/
> MuLan http://www.npl.uiuc.edu/exp/mulan/
> 
> In 2006-7 we will have the main physics production runs, where we 
> clearly will need HPC resources for processing.
> 
> We are planning the following requests from UIUC and UC Berkeley:
> 
> year    project PI               SU        data on Unitree
> 
> 2005    MuCap   Kammel(UIUC)   10000       5 T
> 2005    MuLan   Debevec(UIUC)  10000       5 T
> 2005    MuCap   Freedman(UCB)  10000       same data as Kammel
> 
> 2006    MuCap   Kammel(UIUC)   150000     25 T
> 2006    MuLan   Hertzog(UIUC)  150000     30 T
> 
> The 2005 requests will be Development Applications, the subsequent 
> requests will be full proposals.
> 
> I just submitted the MuCap application after talking to you on the 
> phone.
> 
> Thank you very much.
> 
> 
> Peter Kammel                                David Hertzog
> co-spokesman MuCap                          co-spokesman MuLan
> 

Peter Kammel  /  pkammel@uiuc.edu
Department of Physics, Loomis Laboratory
University of Illinois at Urbana-Champaign
1110 West Green Street, Urbana, IL 61801 
Tel (217) 333-5424 / Fax (217) 333-1215