The growing cost gap between DRAM and storage together with increasing database sizes means that database management systems (DBMSs) now operate with a lower memory to storage size ratio than before. On the other hand, modern DBMSs rely on in-memory search trees (e.g., indexes and filters) to achieve high throughput and low latency. These search trees, however, consume a large... Read More

Lily will present this paper in this meeting: Title: Native Store Extension for SAP HANAConference: VLDB 2019 Read More

You wouldn't run your application on the cloud because it is the cool thing to do. You'd run it there because it makes sense to do so. In this talk you'll hear why some of our customers have moved to the cloud while others have not. You'll also learn how we overhauled Vertica's architecture to become one of the few analytics... Read More

In this talk, we will discuss the challenges of supporting traditional database operations (ingestion, schema design, querying) to the domain of visual data, i.e. images, videos, PDFs, etc. Recent innovations in artificial intelligence have enabled computers to "see" the same structure in visual data as humans. As a result, new systems are emerging which enable users and applications to manage... Read More

This talk is a summary of my soon to be released column in ACM Queue titled "Write Amplification versus Read Perspiration". In this short discussion, we observe that there is a strong pattern in which writing data incurs and obligation to do more work to make it easy to read that data later. We frequently talk about write amplification to... Read More

In this talk we discuss LSM compression for a KV store. In our KV store, we write to an underlying shared storage system that models data as named extents (up to 2GB) and variable-length fragments contained within the extent. Fragments are max of 1MB and are the atomic unit of read and write. Our KV store reads fragments into 64K... Read More

In this talk I will give a sample of some of the research done at AWS. In particular I will talk about some recent results in Reinforcement Learning using a combined on-policy and off-policy approach to obtain rapidly converging and sample efficient algorithms. The key idea in this work is to use propensity scoring and effective sample size reweighting to... Read More

The current research trend is on developing "learned" components to supplement and replace legacy components in database management systems (DBMSs). Such learned components use machine learning (ML) methods to identify non-trivial trends and correlations in the DBMS's runtime behavior. They then use this information to create execution strategies and data structures that are tailored to the application's access patterns. The... Read More

Title: NEVER use mmap for your database Abstract: MMAP can be used as the buffer pool manager for DBMSs. But is it good to use mmap for DBMSs? Chenyao is going to share evaluations for mmap, pitfalls in mmap, and methods to make mmap safe in existing DBMSs. Read More

Non-volatile memory technologies (NVM) enable persistent media to be directly accessed by the CPU through its caches. The biggest challenge introduced by NVM is the little control the application has when persisting data. This stems from the fact that it is not possible to prevent data from being evicted from the CPU cache to NVM at arbitrary points in time,... Read More