TRE sustainability and operations#
Sustainability needs to be long term, but how do you plan for it when the scenario may change in 5 years? There is also an issue with research, this is a service yet funding requires teams to appear to be doing something new each time, and funders often prefer not to pay for infrastructure (also challenges with cost estimates and under/over expenditures).
There are several variables and questions about whether they should be free at point of use (distributing against overheads), or whether to employ a membership user model, a project fee model, standard features being free but charging for high demanding ones or something else. In all cases at least some core funding is required to ensure continuity, specialisation and quality.
What we want to ensure is that a public service exists.
Create a roadmap that focuses on:
Information governance requirements
10 year funding plan
Sustainability from funding perspective beyond the initial 5 years
But what are things going to look like in 5 years time
CL centrally funded model
Service in place, refreshed but need to appear to do something different each time to secure funding.
How costing then? Free at point of use, cost distributed against overheads.
Constrain in the cloud?
Barts recover work space costs from research projects, distributed central cost on a membership/license/user model
Difference between model for internal and external users.
Standard provision free, high storage/compute needs to be recovered
More paperwork to create and chase invoices.
no funders like paying for infrastructure
What counts as core if it was funded?
Duties imposed as data controllers law, or interpretation runs counter to wants of researchers
Folk specialising, if it doesn’t get funded for the future that capability is lost.
Regional SDE model might lead the way of costing-funding-recovery
Some central funding
Specialist areas - operational team
Different environments work differently from researcher perspective
Business and operations to use OS TRE safely and securely
what is the perfect TRE/SDE environment future consolidation
Software development can be amortised across the community
Who provides desk-side support
Tracking usage, egress process, layers of tools and processes that need to be in place
In/out nature of TRE, tiered sensitivity? Commercial sensitivity. Has auditability in the TRE, does it need to be?
Why different for UCL TRE?
Difference in TRE makes funding case easier, adding something new made it more interesting.
Using research funding to backfill
Estimate in advance what project is likely to use, operational costs, usually completely wrong and go over project
Not sustainable to go consistently over budget
Bill after usage is best, but challenging for proposal/funding
Cliff edge, have funding but only sufficient for 1 year not 3 years of project.
Following Access to HPC model
What can you take off the board if problem is solved strategically
Good training for Data scientists: SC like training relevant to disciplines
Seems like we’re trying to boil the ocean
VDI, Excel may be R, Stata
Developing things to deal with core use case
Core capabilities, exceptional stuff is great, but majority, early stage users, standardise and simplify.
Whatever it is, what’s missing the ability to understand data. GIGO
Standardisation of data makes it seem simpler than it is, reproducibility?
AI/ML store data for XX years, is it readable in that time?
Who picks up the storage costs for the data.
How can we make it more transparent
Constrained with the current model.
Guidance provided by RCs, institutional risk as the org have underwritten the project.
This breakout room continued during the second round
Concerned about being able to provide a service, don’t control budgets
Sustainability of providing a public service, rather than generating a business case
SNSDE comes under DH budgets, makes things easier
HDRUK MRC led 20 year vision 5 year cycle
UKBB core underpinning funding
Fund TREs for 3-5 years for specific projects
Specific use cases not currently supported
Individual researchers and work with them and the RO.
Free at the point of use funding?
Provide underpinning capacity?
What is ONS Model?
Free at point of access
Don’t know how the budget is secured
Funding comes through different sources ADR UK
Research proposal, existing staff funding or contracted.
For commercial and public researchers usage has to be for public good, commit to publishing and not for profit
Virtual machines provided some policy for standardising storage/compute available
Trying to enable research
Driven by what researchers ask for
Intrinsic limit on budget call
Budget for a specific network/platform
Leverage external investment
Some Pharma match funding
Universities also fund
Move to long term funding
Strategic level of funding, buffered from long-term budget
Hub large funding but cliff-edged
Free at the point of use
Incentivised-disinsentivised, equity of access
Power users can over-consume, less accountability not having to justify use
consuming data token publication and harvesting data for private use
Free at point of access so data is freely accessible
Reminder: Don’t offer data for commercial use
Ingress-egress labour intensive to pour human eyes
Automation tools for validating statistical disclosure test
Tools and more people-more efficient tools; more people would always be good.
All TREs have these issues, share the solutions
More automation -IDS (Integrated Data Service- SRS Secure Research Service
Free at point of use?? Cuts out some of the applications automated validation of inputs
Understand the whole pathway
Fix one part and it just shows the next bottleneck
Fraunhoffer 1/3-1/3-1/3 lights_on-academic-commercial_activity
Sustainability, prime an initiative without committing to long term investment
More people - more monkeys on typewriters
Over focus on the medical use case currently, needs to rebalance.
Better understanding and economy of scale from small numbers.
Focus critical mass on small number
DARE UK would create a TRE to handle data as an offering
What is a TRE?
At what point does a federated TRE network become a single TRE?
TT: At the point at which you have seamless transition between TREs?
Trust that the analysis/code is running as intended?
What would a solution to this problem look like?
What resources would be needed (people, time, funds, infrastructure etc.)?
How can this community support you in getting them?
What working groups/orgs are already working on this, if any? How can we collaborate with them effectively?
A roadmap should address
Technical knowledge, skills, TRE staff skillsets
Why doing this has to be part of retaining people
Localising staff makes this easier, central models push more to thinking about pay
To address retention
Pipeline of talent
Can TRE model work in R
Not just technical, IG, where can I get more information
Embedded technical/operational/IG knowledge relevant to the problem.
Research - teaching balance.
Lots of politics, in HPC communities, good for those who get it. Not good for those who have to resort to begging
Not necessarily good for SDE
Analysis will follow data
People with data will need to bolt compute
HPC allocation modelled SDE account for compute/storage costs
Why should SDE and HPC be considered differently
10 year plan - scope for accreditation
Chartered research infrastructure?
CSP platform neutral certifications for Data/Cloud