160 likes | 260 Views
User requirements for UK e-Science grid environments. Bruce Beckles University of Cambridge Computing Service. Background (1). Growing technical concern about “grid technologies”: Suitability and Direction particularly amongst “grid developers” and deployers over the last year
E N D
User requirementsforUK e-Science grid environments Bruce Beckles University of Cambridge Computing Service
Background (1) • Growing technical concern about “grid technologies”: • Suitability and Direction • particularly amongst “grid developers” and deployers • over the last year • “Workshop on Requirements Capture for Collaboration in e-Science” • Intense discussion within the UK Grid Engineering Task Force (ETF)
Background (2) – WSRF • The WSRF standard: • announced at the start of the year • “the way forward for grid technology” • WSRF is an “open standard”: • designed “in secret”, with “limited consultation” • by the Globus Alliance, IBM and HP • Concern about the viability of current e-Science projects
A Technical Perspective • “grid technologies” and “grid infrastructure”: • “low level plumbing” for “grid applications” • If the plumbing is wrong then application development is futile. • “The Globus Toolkit” / “WSRF”: • What is the question? • No question, can’t critique!
Another Perspective… Grid technology developer: “Here’s a thing I just developed. I’m sure it’ll be useful to you for something or other. Go on, give it a whirl…” Grid infrastructure developer: “OK, I’ll deploy it so that my application developers can use it. Boy, this sure is complicated to deploy… Umm, what did you say it should be used for again?” Grid application developer: “Right, so this is the latest grid technology? Great. I’ll build it into my application… Now my application is five times as large, doesn’t do anything useful and I have a migraine. Why am I doing this?” End-user: “I am confused. Please help.”
What problem are we trying to solve? • Difficult problem but no clear formulation: • can’t use systematic approach, so… • use combination of: • Intuition: complex problem, intuition unreliable • Trial and error: complex problem, lots of trials! • Luck • The ETF’s goal: To deploy grid infrastructure that is useful to the UK e-Science community • Is it useful?: • Don’t know unless have a clear, technical statement of “useful” • So…
The ETF’s position (then) • Concerned that: • not delivering useful “grid infrastructure” • not clear what it should deliver • not clear what “grid technology” to use • A requirements gathering exercise: • Gather data to help address these concerns • Some more enthusiastic than others
Proposed Requirements Capture Exercise • Discussed on ETF mailing list • Active involvement by ETF members • Intended consultation with: • UK e-Science Usability Task Force (UTF) • UK e-Science User Group
Phase 1: Contacts • Canvass ETF: • opinions on relevant: • issues • people • detailed interviews as appropriate • Ask e-Science Centres via ETF: • description of their projects • contacts for these projects • contacts for other projects that might benefit from “grid technology”. • Examine UK e-Science Stakeholders Meeting attendance list • Use ‘snowball’ technique to gather further contacts
Phase 2: Requirements Elicitation • Analyse contacts to determine ‘representative’ cross-section • Undertake contextual interviews: • about 5 interviews/‘user group’ • about 10 such groups • Raw data made available to interested parties
Phase 3: Requirements Analysis • Data analysis: • Contextual Design (Beyer & Holtzblatt) • Interaction Design (Cooper & Reimann) • Produce: • user personas (‘grid users’) • user scenarios (‘grid application use’) • Derive: • detailed use cases • technical specifications
Phase 4: Validation • Choose representative sample of contacts • Design questionnaire: • Describe newly generated requirements • Ask respondents to rate them in importance • Follow best survey practice
So what happened? “the ETF should not put too much effort into gathering new statements of requirements” “other groups should take this responsibility” – Director, . UK e-Science Core Programme (as reported to the ETF) .
The ETF’s position (now) “Requirements analysis? Never heard of it.” (I paraphrase)
And from this we learnt? Requirements gathering is a political activity. • So either: • be willing to defy political authorities, or • convince them that you will validate their decisions, or • convince them that you will produce nothing of interest.
So now? • No longer directly working for e-Science Programme. • Still working within University of Cambridge • Responsible for: • promotion of e-Science and its related technologies within University • allows a scaled-down version of requirements gathering • University of Cambridge: • leading research university, so… • likely to have a reasonably representative sample of potential users of “grid technology”. • So watch this space… (and all help gratefully received!…)