GAT Meeting - 2022-10-27
- Ann Su
Agenda
- Welcome
- Follow Up Item: Use Case Updates
- Deep Dive: Market Background and Trends
- Are you on trend?
- For the good of the order
- Housekeeping
Refer to the PPT (which can be found here) for additional details on the meeting minutes and discussions.
When available, meeting minutes should be read, and any corrections sent to Maureen.
Participation
Meeting Minutes
#2 Use Case Updates
David Clements gave an update on the use case work. Ed-Fi developed A Practical Guide to Interoperability Use Cases in 2018 which was more industry focused. Ed-Fi also had use cases aligned to dashboards which have now been moved from within Ed-Fi’s portfolio to community support. We are focusing on the dashboard based use cases.
The current use cases are focused on the user persona and what they are trying to accomplish and why. They are being organized into a library structured in big categories, including resource requirements that are linked to API references - what is needed to satisfy the use case.
This gives us an opportunity to discuss what needs to be changed, and what is common across the community and can be updated with new information over time.
Goal is to define the common needs – there’s a lot of use cases across the standards and these need to be curated re what is important to the community.
Working on this to:
- Use on data standard side (such as help within HR and finance data discussion)
- Use it to define and better update certification process based on community needs
- Use it on the data-out side – discuss the usage to handle downstream interoperability (put into action to be kept fresh and move forward)
Will you come back to Use cases that map to different standards – now people get easily confused? Original use cases in the original document are referenced in the new work as well – still exist as use cases.
- Will check on whether this information exists within Project Unicorn.
From an internal perspective this may be helpful to go through – want to play to our strengths and not walk into other organizations’ strengths. Should be able to develop the current work and use with discussions with other standards organizations.
A recent publication may be helpful: Education to Workforce Indicators https://www.mathematica.org/publications/education-to-workforce-indicator-framework-using-data-to-promote-equity-and-economic-security#:~:text=The%20Education%2Dto%2DWorkforce%20Indicator,K%2Dto%2Dworkforce%20continuum. This is a reference tool that creates a connection across PW20 spectrum.
Part of the Community of Education is working to publish information on and leverage communication with different standards body. CCSSO is a good third-party participant in this space for recommendations on standards organizations working together.
#3 Deep Dive Discussion: Market Background and Trends
Eric and David have been on a listening tour with school districts on what has and hasn’t worked on last several years and industry fact-finding activities and found these themes of change:
Highlighting where we were in 2015 – data warehousing projects that had large front-end costs and expensive to maintain.
- Many of these older projects failed – difficult to organize, costly. Today’s cloud offerings have a lot of value and allow districts to start small with a single problem, get it working and see what is valuable. Actual cost for starting small is more in line per student than traditional warehousing projects. Old projects were difficult to maintain, and succession plans were difficult – IT staff turnover and project was in jeopardy.
- Key benefits of Cloud approach:
- Cloud-based – zero upfront infrastructure
- Pay-for-usage pricing models
- Alternatives to complex ETLs
- Data governance efforts are aligned with value created
- Minimal skills to get started + training programs
- Software costs likely in the $1-2 per student FTE range annually
- Shift to being applied more broadly to analytics space.
- Moving forward districts installing own ODS are unlikely - will see much more involvement with managed service providers doing this for them.
Does this resonate with GAT members? Any additional context or experience.
- LEA – like this idea but it does not reflect my realty – we don’t have anything in the cloud and have our own data warehouse that we built and concern re being able to shift that into a space of not relying on internal personnel; actual cost of maintaining across the systems is much more than $1 per student. There is not a single point of failure – based on staff overturn not able to maintain as wanted – haven’t replaced former staff who worked on putting this in. Also dealing with two internal contractors who are maintaining the warehouse. Pushback on getting ready for 4.0. Thought we were sustaining systems we have but that may not be the case. Board believes anything in the cloud is susceptible to data being stolen – even after multiple briefings.
- LEA - Echoing everything said by other LEA. The people who maintain our ODS have not left yet but very concerned about that future turnover. Are not able to offer salaries that allow us to hire what is needed. Also, additional education funding is going away in two years. Will be a huge fiscal cliff.
Other comments:
- These are good points. So although it may be "minimal skills to get started" it doesn't necessarily mean it's sustainable… sustainable in the context of the "single point of failure" when resources leave, costs, etc.
- Increasingly there will be talent pipelines around particular cloud platforms and the investment in that should bolster our ability to maintain education systems over time.
- Public procurement point is not just the cloud security but also capital vs operating perspective – financing will give us capital funding but not sustaining $$. Sustainable in the context of the "single point of failure" when resources leave, costs, etc.
This shift will impact Ed-Fi tools and sustainability.
Any other trends as we discuss what 2023 should look like? How can we encourage and seed sharing across the community? Are data lakes “dead”?
- Data lakes not dead. For example, Tennessee struggling with doing Ed-Fi since 2013 – they have a different solution that the back end is not an ODS but their own data lake using the Ed-Fi standard and are going to try and remove some of the extensions they created. PS has tried to convince them to get full behind Ed-Fi but still running into lots of problems. It is a struggle internally because they don’t want to stand up an Ed-Fi ODS and they render databases in their own sequel. We need to acknowledge there will be other ways than sequel server ++. Trend is looking for scale to enterprise – would love for PS to put Ed-Fi API in front of all our tools. It is a way to hook up Ed Fi to a back-end enterprise – could be data lake or something else.
- One of our goals for next year is to test other emerging technical solutions that may not be the same architecture as the past.
- IN still needs to solve for consolidating current year (ODS) and longitudinal data (warehouse) and other agency data to a single data source for application consumption. Data lake is at least a consideration.
- Depth is a real concern. Cloud does not equal easier, in our experience, but it does provide us an opportunity to share responsibility in a different way.
- From higher ed – continue to be challenged as teacher prep being part of a larger university IT system; the intersection is that anything that is lighter weight and requires them to build and maintain fewer things locally are good – don’t have a lot of money and resources for their system as one small part of larger IT system.
- Shout out for the hard work and dedication of Ed-Fi Alliance and GAT to focus on and make happen continued enhancements to Ed-Fi over the past ten years!
Action Item
- Use Case Update - Will check on whether this information exists within Project Unicorn.