![]() |
![]() |
| Time | Activity | Location |
| 3:00pm - 6:00pm | Registration begins | Hearst Social Hall |
| 6:00pm - 7:30pm | Dinner | Crocker Dining Hall |
| 7:30pm - 9:30pm | Refreshments (Soda, Beer, Wine, Snacks) and Chat | Chapel |
| 9:30pm - Midnight | Night Owls' | Chapel until 11pm and later transitioning to Director's Cottage |
|
* Breakfast, Lunch and Dinner are held in Crocker Dining Hall. All talks are in the Chapel. All Night Owl's events are in Chapel until 11pm and later transitioning to the Director's Cottage. |
||
| Time | Activity/Subject |
| 7:30am - 8:30am | Breakfast |
| 8:30am - 8:40am | Opening Remarks - Ippo and Justin |
| 8:40am - 10:00am | Session 1: Keynote and Student Intro Chair: Ippokratis Pandis |
|
[1 hour] Keynote: Title TBD - Ion Stoica (UC Berkeley) [20mins] Student Intros - Ippokratis Pandis |
|
| 10:00am - 10:30am | Break and Snacks |
| 10:30am - 12:00pm | Session 2: Putting database researchers out of a job Chair: Marc Brooker |
|
[20 mins] The Future Is Bespoke: Synthesizing One-Size-Fits-One DBMSs with LLM Coding Agents - Carsten Binnig (TU Darmstadt) [20 mins] Towards AI-Generated Database Management Systems - Jiale Lao (Cornell) [20 mins] Specification-Driven Synthesis of Query Execution Logic for Novel Storage Systems - Puneet Mehrotra (University of British Columbia) [20 mins] Learned Systems Are Dead; Long Live Learned Systems - Todd Lipcon (Google) |
|
| 12:00pm - 1:30pm | Lunch and Break |
| 1:30pm - 3:00pm | Session 3: What data stack do our agents need? Chair: Nathan Bronson |
|
[20 mins] Execution LogAct: Enabling Agentic Reliability via Shared Logs - Mahesh Balakrishnan (Meta) [20 mins] The Postgres Trap: Database Incrementalism in the Age of Agents - Ajay Rayasam (Osage University Partners) [20 mins] Commits, Transactions, Durable Execution:Git-for-everything and the self-driving data stack - Jacopo Tagliabue (Bauplan) [20 mins] AI for Production - Mayank Agarwal (Resolve AI) |
|
| 3:00pm - 3:30pm | Break and Snacks |
| 3:30pm - 5:00pm | Session 4: Architecture Lessons Chair: Sailesh Krishamurthy |
|
[20 mins] Ten Years of Magic Pocket: Lessons From Scaling an On-Prem Exabyte-Scale Blob Store - Facundo Agriel (Dropbox) [20 mins] How We Built Spanner Omni - Sebastian Kanthak (Google) [40 mins] Preventing and Mitigating Logical Corruptions in Salesforce Database - Vaibav Arora & Jameison Martin (Salesforce) | |
| 5:00pm - 6:00pm | Break / Walk on Beach |
| 6:00pm - 7:30pm | Dinner |
| 7:30pm - 9:00pm | Gong Show Emcee: Shel Finkelstein |
|
[5 mins] What Postgres would look like if we started today - Manish Jain [5 mins] Your database is all you need - Qian Li (DBOS) [5 mins] From Vitess to Multigres: Lessons learned - Sugu Sougoumarane (Supabase) [5 mins] Generative AI; Generative Testing - Kyle Kingsbury (Jepsen) [5 mins] Resurrecting Semantic Concurrency Control - William Schultz (MongoDB) [5 mins] Open Storage Format for OLTP (Why is it difficult, is it possible, what doors will it open) - Sunkanmi Adewumi (UC Berkeley) [5 mins] Scalable Image AI via Self-Designing Storage - Utku Sirin (Harvard University) [5 mins] Bandwidth Is the Bottleneck: Rethinking Scale-Up Analytics for CXL - Georgiy Lebedev (EPFL) [5 mins] A Little Little's Law Keeps the Metastable Failure Away - Kiran Kumar Muniswamy Reddy (MongoDB) [5 mins] Failure resilience in large-scale LLM training on TPUs - Alexander Shraer (Google) [5 mins] LLMs are compute. Context is memory. Inference speed is king. - Mehul Shah (Glean) | |
| 9:00pm - Midnight | Night Owls Session - (Chapel until 11pm and later transitioning to Director's Cottage) |
| 7:30am - 8:30am | Breakfast |
| 8:30am - 10:00am | Session 5: Hardware will save us all Chair: Natassa Ailamaki |
|
[20 mins] The Fastest Way to be the Fastest - Bobbi Yogatama (NVIDIA) [20 mins] The Next Watt Saved Won’t Come from a CPU or a GPU - Devashish Purandare (Microsoft) [20 mins] What Changes When Memory is Disaggregated? Building a Database to Find Out - Chiradip Mandal (Space-RF.ORG) [20 mins] Using SmartNICs for Database Management - Phil Bernstein (Microsoft Research) |
|
| 10:00am - 10:30am | Break and Snacks |
| 10:30am - 12:00pm | Session 6: Query Processing Chair: Peter Boncz |
|
[30 mins] CoddSpeed! Hardware Accelerated Analytics in Microsoft Fabric - Matteo Interlandi (Microsoft) [30 mins] Privacy Policy of Cloud Data Services is Harmful - Surajit Chaudhuri (Microsoft) [30 mins] ROII-Guided System Design - Natassa Ailamaki (EPFL) |
|
| 12:00pm - 1:30pm | Lunch and Break |
| 1:30pm - 3:00pm | Session 7: Object Storage and Open Formats Chair: Rebecca Taft |
|
[20 mins] Beyond Disaggregated Storage: The Case for Open Storage in OLTP Systems - Sergey Melnik (Databricks) [20 mins] Ex Falso SQL Quodlibet: Excluding the Middle - Chris Douglas (UC Berkeley) [20 mins] S3 is the Future, S3 is the Past - Viktor Leis (TU Munich) [20 mins] Operational Benefits of Databases Built on Object Storage - Colin Breck (Tesla) |
|
| 3:00pm - 3:30pm | Break and Snacks |
| 3:30pm - 5:00pm | Session 8: Failures Considered Harmful Chair: Atul Adya |
|
[20 mins] Real Life is Uncertain. Consensus Should Be Too! - Reggie Frank (UC Berkeley) [20 mins] Advancing Storage Efficiency in Replicated Databases - Scott Andreas (Apple) [20 mins] Retry is an open problem - Peter Alvaro (UC Santa Cruz and Amazon Web Services) [20 mins] MESSI Reasoning about Performance and Metastability - Aleksey Charapko (University of New Hampshire) |
|
| 5:00pm - 6:00pm | Break / Walk on Beach |
| 6:00pm - 7:30pm | Dinner + HPTS Business |
| 7:30pm - 9:00pm | Panel followed by Gray-Beard Poster Session - Chair: Natacha Crooks |
|
Panel - AI and Systems:
Todd Green (Microsoft) Randy Shoup (Thrive) Mark Little (IBM) Gray-Beard Poster Session |
|
| 9:00pm - Midnight | Night Owls Session - (Chapel until 11pm and later transitioning to Director's Cottage) |
| 7:30am - 8:30am | Breakfast |
| 8:30am - 10:00am | Session 9: Distributed Systems Are Hard! Chair: Justin Levandoski |
|
[20 mins] Please Implement A Quorum Database - Alex Miller (Oracle) [20 mins] Til Depth Do Us Part: Commitments and Determination - Joseph Hellerstein (UC Berkeley and Amazon Web Services) [20 mins] Durable Execution: An End Run on the End-to-End Principle? - Johann M. Schleier-Smith (Temporal Technologies) [20 mins] Autoscaling Is Not a Resource Allocation Problem - Sacheendra Talluri (Clickhouse) |
|
| 10:00am - 10:30am | Break and Snacks |
| 10:30am - 12:00pm | Session 10: Lakehouses vs HTAP: The showdown Chair: Natacha Crooks |
|
[40 mins] Lakebases & Lakehouses => LTAP - Ippokratis Pandis (Databricks) [20 mins] Lakehouses don't cut it (yet) - Stefano Stefani (Goldman Sachs) [20 mins] No, HTAP in the Cloud Is Not Solved - The Industry Traded Freshness for Scalability and Called It Innovation - Daniel Ritter (SAP) |
|
| 12:00pm | Farewell |
| 12:05pm - 1:30pm | Lunch |
Program Chairs |
|
|
Natacha Crooks (UC Berkeley) |
|
Program Committee |
|
| Atul Adya (Databricks) | Peter Boncz (CWI) |
| Nathan Bronson (OpenAI) | Marc Brooker (Amazon) |
| Alon Halevy (Google) | Stratos Idreos (Harvard University) |
| Fatma Ozcan (Google) | Deepti Raghavan (Brown University) |
| Xiangyao Yu (University of Wisconsin - Madison) | Irene Zhang (Microsoft Research) |
Organizing Committee | |
| Ippokratis Pandis (Databricks) | Justin Levandoski (hiddenweights) |
| Shel Finkelstein (Independent) | Pat Helland (Salesforce) |