Notes for Dec 2024 Virtual ACCESS Quarterly Meeting
Link to Google Doc for original notes:
https://docs.google.com/document/d/1QSLjzsgSuKWZWgeLx9aHYBHGnr8pMeJvbYyU423MOPg/edit?tab=t.0
What’s Everyone Developing
Google Folder: ACCESS "What's everyone developing" Dec 2024
Meeting information
The purpose of this meeting is for all ACCESS teams to share what they are currently working on or are planning to start to work on in the next quarter.
Agenda
Time (EST) | Topic | Presenter |
11:00 - 11:05 | Introduction | Joe White |
11:05 - 11:30 | Allocations Team Updates | Matt Yoder & Nathan Tolbert |
11:30 - 11:55 | Operations Team Updates | JP Navarro, Derek Simmel, Dave Wheeler |
11:55 - 12:00 | Break | - |
12:00 - 12:25 | Metrics Team Updates | Greg Dean & Doug Southworth |
12:25 - 12:50 | Support Team Updates | Alan Chalker at al. |
12:50 - 13:00 | Open Discussion | Joe White |
- 1 What’s Everyone Developing
- 1.1 Meeting information
- 1.2 Agenda
- 1.3 Meeting Notes
- 1.3.1 Allocation team update
- 1.3.2 Operations team update
- 1.3.3 Metrics
- 1.3.4 Support
- 1.4 Open Discussion?
Meeting Notes
Google Folder to store presentations: https://drive.google.com/drive/folders/1VpSCYhe1O74A_FCX9BpBRrLInNelELMD
Dec 2024 (Virtual) ACCESS Quarterly Meeting
Allocation team update
Presenters: Matt Yoder & Nathan Tolbert
XRAS
Nathan -
Much will be items covered in the last meeting but we will be providing status updates.
Focusing on efforts that impact other teams
On-Ramps - widget that lives on other partner websites
RMACC is now live - Illinois Compute and NewMexico will soon be next
Don’t need a lot of features requested - more what is there and how to start
Resource Catalog
Variable Marketplace - working on user interface and all the plumbing
Goal was beta release at end of year - will be moving to January
Moving resource controls that are allocation specific to XRAS Admin
Provides better interface and variable market controls
Class Registration - with Support
Allocation worked on backend
Goal is making it easier to add in groups of people for classes
Q - User Profiles and people not filling in NSF user status - how to not miss information
Support will talk about this in their session
AMIE Multi-federation for NAIRR
Still in progress - exploring solutions - no timeline yet
Q - variable market place - how does this play into what PIs see when they apply - will RPs be advertising?
A - if there is a discount there will be banners
B - not sure how much actual promotion we will do vs what RPs do
C - maybe a carousel of deals
Q - Puri - NSF can’t log in - Lavanya will fix
Matt Yoder
Resource Catalog
This is not just an allocations effort
Review of history
Working on implementing next Monday
Almost ready to share!
Worked on data prep and publishing and have had user testing
RPs reviewed Beta in November
Resource Data
15 AP endpoints from all teams
Identified common fields for Resource and Groups - so all can be pulled
And combined in Browser
Catalog Interface - a few things have changed
Removed carousel - replaced with pathways typically needed
Detail page - simplified language of headings - not familiar with terms
Recent projects was changed to examples of papers of how others are using the resources
Explained more about Wait Time
Bug Fixing and improvements
2 pending changes before launch
5 program resources are not in new catalog
Add breadcrumbs to match ACO Site Navigation
Thank you to all who worked on this.
Q - How to announce to the community that it is available? And are we collecting feedback?
Have a call to create blurb in January newsletter
Need suggestions on how to collect feedback
Andrew P will have something - but won’t be ready for January
They have a standard form
Have a call or meeting with everyone who is interested in feedback - in January
Comms meeting this Thursday - they will be talking about this
Will let PM know if they need help on getting feedback
Q - what is plan for when an endpoint goes down - how is it being monitored? With that many APIs there is bound to have one down regularly
JP suggests - work with operations to get the right queues to submit issues or suggestions to the catalog
Cyprus might be able to help with this
Operations has a Nagios that could be used as well
They will follow up on this
Operations team update
- JP Navarro, Dave, Cathy, and Eric
JP Navarro
Current projects
Resource Integration Badges - almost to Beta Phase
Developing resource and staff interface
They created http://dashboard.operations.access-ci.org - internal only
Worked with all ACCESS tracks
Upcoming
RP and staff user beta phase 1 in January
3 phases of rollout - improvements and new features
2 includes updating badges
3 includes adding new resource
Hope to be done at end of PY3
Infrastructure is looking at Regional Network and Data Collection integration
Resource Information
Resource Groups and Group Types in CiDeR - done
Used to be in an excel sheet - this was copied and transferred over
Just go straight to Cider for the information
This is where they need to maintain it
RG description APIs - done
Resource Group New API - done
RP Software publishing, curation and SDS
Planning
Badges APIs
Integration Badges in Resource Catalog 2
Caches …….
Resource Groups and Group Types affect all the tracks
Reviewed a hierarchy diagram in the presentation
System status news - can get news from central services for outages
Q - Regarding pushing notification to ACCESS users, the current news posting will only be sent to subscribers. Will any of the current support related efforts resolve this issue.(e.g. sending emails to all users from one RP)?
Andrew Answer: Affinity Groups can send email to all people with an allocation with an RP. The Support Announcements are sent to all people with allocations at any RP. The caveat is that people can remove themselves from this list. Perhaps the wording on that option should be updated? There has been some discussion about another list that would not allow people to opt out.
Guangzhen Jin - Appreciate the options @Andrew Pasquale (he/him). Those would still serve as workarounds for us.
Stephen D - Even if you DID have a list you couldn’t opt out of, you still face the issues of:
People not reading their mail
People creating filters
This has been asked for - we should do this - need to negotiate who should do this - Support or Allocations or a combination of teams.
Follow Up - needs more discussion - short meeting with Support, Operations, Allocations and any other track - how to email users - JP will work with Shannon to schedule meeting 30 mins.
Discover and reviews Access/Science Gateways
Resource contacts maintained in CiDeR - done
Will be creating views for RPs and staff - provide access to all emails for use
Suggested Work
Status for Virtual Host - Syslog feeds into Omni-SOC - under consideration
Derek Simmel
OmniSOC - Security Information and Event Management
Working on allowing other tracks to submit to OmniSOC
Reviewing instructions for RPs to use this service
SSH Pubkey Service - in production
a backend update to the co manage is requiring us to roll out and test a revised version of their database
When we have it’s ready, we will let folks be aware and invite RPs to try out and working as expected
Active development in incident response procedures
Asking RPs to participate in exercise
Dave Wheeler
Network peering with FABRIC - almost ready to send to EC
Sending Network Data to NetSage RPs as well as us
Continuing to encourage RPs to send updates to NetSage
Data Transfer Application Evaluation
Evaluating data transfer alternatives
JP -
Policy document exists for data transfers - shared with resource provider forum - recommended it be adopted by RPs as well
Includes when you should notify re: service status changes
And How you will notify
Relies on many teams - effort between operations and support
BREAK
Metrics
XDMoD Development
Ver 11.0 released
Working on fixes for bugs related to that
Including ACCESS Credit Equivalents for Cloud (Jetstream2)
Integrating job level performance data from Delta and ACES
Improvements to XDMOD training data with support AI Chatbot
Generating new training data with non-technical queries to help with conversational queries
Starting work on mechanism for embedding users job performance data from xdmod into OnDemand instances
Open XDMoD Development
Released ver 11.0
Swapped HighCharts for Plotly
Support Rocky Linux 8
New resource specifications realm
Ver 11.5
Support for Rocky linux 9
Support for multiple org for resources and people
Data Analytics Framework
Hosted
Bug fixes
On-click access - no additional login
Browser based
Functioning prototype using JupyterHub
Software Dec
Changes to support Cyber Simulator
Feature requests
Bug Fixes
Analyze network data from Sage
Workload Analysis Reporting
Reports to NSF senior leadership
Supply and demand
ACCESS + wider ecosystem
Demographics
Future
Auto gen quarterly reports
Download queries via portal tab
Reports for RPs and Training Events
Template showing usage and performance information
Template report for PIs running training events
CloudBank Integration
Current status
Implemented NAIRR specific data flow
Doing final testing - should be done by eOY
Future
Add PI Org by Group
Add Fund and Account
Work on Compute Usage
NAIRR Pilot
Making a basic instance
Allocations
Public Cloud
Academic Cyberinfrastructure (Frontera, Lonestar6, ACCESS/Bridges/Delta
Initial reports
Latency between award and usage
Projects with no usage
PATh
Modules to
Display PATh data
Installed and tested by team - working well
Ingest historical records
Integrating MMS team reports to NSF
Doug Southworth
Netsage - overview of what it is
There is a resource netsage.access.io
We are ingesting Globus data now - every 24 hrs
Can be made to look like flow data - looks like completed tasks now
Navigation is same
Can do confidential tasks as well
Walkthrough
Will be scraping external database information soon and placing in science registry
Q how many sources are in NetSage
A - 6 different regions plus 6 main institutions - soon will bring in tier 2 sites
Q - how researcher facing is NetSage
A - designed to be easy to navigate - so you don’t need to be as technical to use - make it transparent to use
Q - are there interfaces we could point researchers at?
A - Access Dashboard would be first to recommend - other than that we don’t have anything specifically built for that. The questions on the first page help direct user to look at what they want to know.
Support
- Alan Chalker Andres Pasquale
Andrew
Summary of Current Projects
Summary of Upcoming Projects
OnDemand
Current
Q&A tool - on access ecosystem
Over 1000 Questions answered
Will review to make sure answers are correct
Potential collaboration with Operations
Looking at what data and how often to update
Training Registration
Approve or deny
Waitlist
Connect to allocation process to automatically add
Report of registrants
Upcoming features
Affinity group
Pre and post surveys
Metrics (SUs consumed
Planning document link in slide deck
ACCESS Widget - Q&A Bot
Options for opening tickets and providing feedback
Support for Affinity Groups
Work with Coordinators
Identify community needs
Improve documentation
Provide community info
Add features to support community engagement
Software Documentation Service
API currently managed by support used fy the resource calalog to list software installed on Resources
ACCESS Resource Advisor - updates to UI/UX, switch to new software list, new version based on personas
Q - Joe - ARA - when looking at resource allocation - new resources that come online that don’t see high utilization in their early life - make use of recent historical info to promote to the top of the list to get people interested in using them. Can provide that info.
A - Andrew likes this idea
Measuring Topics of Interest
What do they need when they come to our website
Analytics
Knowledgebase
Access Video
Updating and making new videos
Working on a process for creating and reviewing content
Alan Chalker
OnDemand - made great progress on metrics for OnDemand
Most resources have both OnDemand as well as XDMoD
Showed stats for OnDemand Oct 24
Many available stats - examples in powerpoint
This is generally available now
Emily Moffat Sadeghi - Goal is a major release before EOY
4.0 Updates
Walkthrough of what the above changes look like
Open Discussion?
Suggestion - we make updates like this should be 2 hours at future quarterlies
Stephen - one of the best meetings he has been on and really good questions and interaction.
Great job everyone!