re AOB:
Duncan asked about voms dpm issue and status - does it apply to Sl5? 
Yes SL4 and SL5. Should be fixed in next certified release of DPM (1.8.0) which we plan to test 
Sam will contact JP re 1.80 this week but it should be possible to install the new voms on top of an older release if one wanted.

CHE PSummary:
DPM side meeting: 
we'll be testing  3rd party rfio and NFS 4.1 
DPM Instumentation:
gathering probes from others via a new contrib SVN  on the dpm site.
Possibly putting our toolkit in there too.

Summary of CHEP: 
Sam will give a talk at HepSysMan on this.
Taiepi was rainy and hot.
Themes: Cloud things - little discussion of IO issues with that.
After Amsterdam storage meeting talk of open protcols: webdev, http acess, NFS 4.1 access.

Dcache nfs4 talk - beat dcap on every test they did - possibly read ahead issues.

xrootd also pushed - particularly glocal redirector , and Brian demonstrated his ability to access data from anywhere with a little latency but good bandwidth. 

Noone demonstrated secure http or webdev interfaces.

Talks from new new expreiments - showed some scarily high data rates.

SSDs are rubbish according some crazy person.

Caching models again post the amsterdam workshop.

Talks on data preservation - archving for the future e.g on Hera-B and Babar.

CernVM-Fs - mentioned several times - reducing IO load for software areas - keenness detected.
(Chris W - also keen)
Duncan pointed out that cernvm fs was not really storage and that maybe this group should widen its remit to be a more general technical advisory one. 

Non storage topics: a lot on many cores and GPUS and how to make any use of them.

Qs for Sam:
Ewan: Central management of software areas from CVMFS?, Sam: yes works for T3s on ATLAS. 
Brian: Secutiry audit ongoing, Sam: Yes .
Sam: git like - reduces the number of stats etc, and amount of storage required. 
Chris W: Reduces significantly load on the metadata server. All their nodes are using it for ATLAS 
Ewan would like a practical talk on cvmfs at hepsysman - Pete: there will be one.

Wahid - other CHEP talks:
Hammercloud - lhcb and cms are now using it - other VOs can too (soon) with plugins. 
EOS - a cern devloped fileystem for the future. Quite some comment on whether we ought to be developing yet another new thing but maybe it will be good and useful for other sites.

GDB 2 weeks ago: 
Some progress on some demostrators - but not that much on any that weren't already planned or underway.
Brian commented that the speedySRM demonstrator had been tested on dcache and whether it should be pushed for other srms was open.

Wahid suggested that next week we should try and pull together the key items from CHEP and the gdb and decide what we wanted to focus on for contributing to (or keeping an eye on). People seemed to agree with that.

Chat window:
[10:02:50] James Thorne Yes
[10:15:52] Brian Davies sorry i'm late
[10:23:59] Duncan Rand chris are all your nodes using cernvm-fs now?
[10:24:31] James Thorne Do you want me to ask Ian to talk?
[10:26:46] Ewan Mac Mahon It's how you store the software b 
[10:30:23] Duncan Rand could someone post the link to the HC talk plese
[10:31:03] Duncan Rand found it
[10:31:53] Wahid Bhimji EOS talk: 
[10:33:13] Wahid Bhimji http://117.103.105.177/MaKaC/contributionDisplay.py?contribId=313&sessionId=118&confId=3