NDSA:Wednesday, April 14, 2011

NDSA Infrastructure call, April 14, 2011

• On the call:
• Karen Cariani
• Dan Dodge
• Trevor Owens
• Mike Giarlo
• Gene Mopsik
• Gene Hurr
• Cory Snavely
• Micah Altman
• John Spencer
• Elizabeth Perkes
• Dean Farrell
• Matt Guzzi
• John Unsworth
• Kurt Ballford
• Andrew Woods
• Mike Giarrlo
• Mike Smorul

(apologies for anyone missed, please add anyone who was missed)

Next Actions

1. Confirm minor edits to the implementer questions. Respond to these on the list.
2. Everyone on the call will work to respond to the implementer questions for their org before the May call.
3. Micah Altmen will contact individuals at Amazon and Google about additional cloud presentations
4. Still need to identify someone to contact someone with Azure

Discussion of Implementer Questions

The call started with a review of Mike Giarlo’s responses to the implementer questions for Penn State. In general, the group felt that the questions were addressing issues we are interested in exploring. The group suggested a few ways to tweak the questions and agreed on a general strategy for keeping track of the questions.

Depth of responses to questions

Several of the participants representing commercial members noted that they may need to give fuzzy answers to some of these questions. After discussing this, there was general consensus that each individual should fill out the questions to the level of clarity that they and their organizations are comfortable with. Related to this, in a later part of the discussion there was general consensus that we would not be publicly sharing these individual responses. These will continue to only be accessible to other members. We will at this info and publish what we think best documents best, or emerging practices and common needs or desires.

If necessary, adapt the questions to make sense for your particular member organization

Several individuals on the call represent membership organizations, or in some way are not directly involved in these storage decisions. There was general consensus that these individuals should adapt the questions in ways that make sense to their situation. In many cases this will involve responding to the questions as general outlines of how their members would respond.

Getting at the “Why” component of several of the questions

The group identified four of the questions as questions that we would like to get at more of the “Why” of the question and less of the “What.” The following represent a first attempt at adding a why component to each of the questions we discussed. Please feel free to edit and refine the questions on the wiki or discuss them on the list.

• 2.What large scale storage or cloud technologies are you using to meet that challenge? Further, why did you choose these particular technologies?
• 5. What are your performance requirements? Further, why are these your particular requirements?
• 6. What storage media have you elected to use? (Disk, Tape, etc) Further, why did you choose these particular media?
• 10. How frequently do you migrate from one system to another? Further, what is it that prompts you to make these migrations?

Storage Systems and the Stack

Discussion of questions 11, 12, and 13 revealed that there is some ambiguity about what exactly constitutes a “system.” The group discussed the idea of expanding the questions to consider the entire preservation stack, but decide that this would become too broad. To this end, the group seemed to feel that keeping these questions but noting that we are interested in the storage component of groups preservation stack to make sure we were still focused on a coherent area. Thinking about this as an examination of the storage component of the preservation stack is intended to walk the line between considering the entire stack and simply considering which storage medium individuals are using. To this end, for questions 11-13 respondents should note if elements referred to in individual questions are handled outside the storage portion of their stack, but should not feel the need to give a detailed description of their entire stack.

There was a suggestion to think about “system” in these questions as “the thing you bought from a vendor –that thing that your servers or management system uses on the back end. The idea in this case is to restrict answers to that so we can determine what can be bought and can’t be bought.

Plan for sharing and dissemination

1. The raw responses will be kept behind the login on the NDSA wiki. So, NDSA members can review them but they will not be made public in their current state.
2. We will then uses these responses as the basis to put together a report on existing and emerging practices. We may well consider a case study approach that focuses on representative examples or case studies from members.

Plan for Implementer questions

1. Before the next call we all review and respond to the questions and post those responses to the wiki. If you are not directly making these decisions try to answer them in a way that reflects the way individuals in your group or association think about them. To that end just adapt the questions when needed.
2. Send out to NDSA list in general after our own WG have answered and filled it out. Make it look like they are adding to something.
3. Consider circulating to targeted un-represented groups after we have finished exploring this inside the NDSA.

Continue to Invite Talks from Cloud Service Providers

While moving ahead with responding to these questions we would like to also arrange for a series of presentations on other commercial services. The group discussed Azure, Amazon, Google Apps and Rack Space. Micah Altmen offered to reach out to contacts at Google and Amazon. We still need to identify an individual to contact someone with Azure.