Hi
Just to remind sites to send replies to the questions to me in the next
day or two so I can collate for the workshop this weekend.
thanks
Duncan
On 29/09/16 13:49, Duncan Rand wrote:
> Hi Jeremy
>
> Yes, I can do that. I already have one response, please could other
> sites send me their replies.
>
> cheers
> Duncan
>
> On 28/09/2016 20:27, Jeremy Coles wrote:
>> Dear All,
>>
>> As I cannot see to whom this message has gone I thought it worth
>> forwarding to the list. It would be useful to compile a GridPP response
>> between those who are most interested in the questions raised. Duncan,
>> perhaps this is something you may like to lead please?
>>
>> Many thanks,
>> Jeremy
>>
>>
>>
>>> Begin forwarded message:
>>>
>>> *From: *Tony Cass <[log in to unmask] <mailto:[log in to unmask]>>
>>> *Subject: **Collecting input for the networking session at the WLCG
>>> workshop*
>>> *Date: *28 September 2016 20:12:58 GMT+1
>>> *Cc: *Shawn Mc Kee <[log in to unmask] <mailto:[log in to unmask]>>,
>>> Edoardo Martelli <[log in to unmask]
>>> <mailto:[log in to unmask]>>, Marian Babik <[log in to unmask]
>>> <mailto:[log in to unmask]>>
>>>
>>> Dear Colleagues and WLCG Workshop attendees,
>>>
>>> To help encourage productive discussions during the networking
>>> session, we’d like to circulate some thoughts and questions
>>> beforehand. On the thoughts side, we see networking issues on varying
>>> time horizons as:
>>>
>>> - Short term (the rest of Run2): can we cope with the extra
>>> load due to the better than expected LHC performance? Can we be IPv6
>>> ready? Can we monitor, diagnose and manage our networks sufficiently
>>> for HEP needs?
>>> - Medium term (LS2/Run 3): Upgrading the networks for Run 3;
>>> moving towards complete availability of data over IPv6. How do we
>>> integrate Software Defined Networking capabilities that exist in our
>>> networks to support our needs?
>>> - Long term (HL-LHC): Do we influence the international
>>> network environment or live in an environment built for more demanding
>>> customers? To frame this another way, will HEP continue to enjoy
>>> abundant bandwidth and prioritised attention or will we need to adapt
>>> to an environment filled with other network intensive science domains
>>> competing for the network?
>>>
>>> In the background there is also the issue of integrating commercial
>>> cloud resources into our networking infrastructure. For this, though,
>>> we imagine that you (and especially the experiments) just want the
>>> network engineers to “make it work” rather than to have any detailed
>>> discussion of the issues at the workshop.
>>>
>>> With this in mind, we’ve scheduled talks on IPv6, network technology
>>> evolution and LHCOPN/LHCONE status.
>>>
>>> In terms of questions, we would like to ask
>>>
>>> Sites
>>> ====
>>> - What fraction of your LAN capacity is used? What fraction
>>> of your WAN capacity?
>>> - On what timescale will you upgrade your LAN? Your WAN?
>>> - If you don’t have an upgrade planned, what are you waiting
>>> for?
>>> o Money?
>>> o Increased traffic?
>>> o Guidance?
>>>
>>> Experiments
>>> ==========
>>> - What bandwidths (LAN & WAN) should Tier-1 and Tier-2 sites
>>> be targeting for Run3? HL-LHC?
>>> - Do you want Tier2 sites to separate clearly into ones with
>>> disk and ones without? Will networking infrastructure influence the
>>> choice or have to be adapted to a choice that is made for other reasons?
>>> - Will there be an experiment focused update of the Bos/Fisk
>>> networking document (http://lhcone.web.cern.ch/node/2935)? If so, when?
>>>
>>> Both
>>> ====
>>> - What obstacles do you see preventing you migrating to an
>>> IPv6 only environment?
>>>
>>> Please send answers to these questions as well as any comments or
>>> thoughts to us ([log in to unmask]
>>> <mailto:[log in to unmask]>, [log in to unmask]
>>> <mailto:[log in to unmask]>) by noon Geneva time on Thursday 6^th
>>> October and we’ll use the input to organise the session. Bonus! If we
>>> get lots of input, you won’t have to sit through presentations of
>>> material available to read on Indico!
>>>
>>> Regards — and apologies if you receive two copies of this mail,
>>> Tony & Shawn
>>> Network Session co-chairs
>>
|