Kai Yu

Kai_Yu_ACED 300x300

Kai Yu

Distinguished Engineer

Dell Technologies, Inc

Biography

Kai Yu is a Distinguished Engineer in Dell EMC Solutions Engineering and a member of Dell Technical Leadership Community. Kai has 26 years’ experience of architecting and implementing various IT solutions by specializing in Oracle RAC database, Oracle BI and Oracle EBS, Virtualization/Cloud. Kai has been a frequent speaker at various IT/Oracle conferences with over 180 presentations in more than 20 countries. He also authored 35 articles in technical journals such as IEEE Transactions on Big Data, IOUG Select and Dell Power Solutions and co-authored Apress book “Expert Oracle RAC12c”. Kai has taken some leadership positions in IOUG such as IOUG conference committee member, IOUG RAC SIG president and IOUG Cloud SIG co-founder and the current vice president. Kai has been an Oracle ACE Director since 2010 and was featured as Oracle ACE Spotlight by OTN. He also received the 2011 OAUG Innovator of Year award the 2012 Oracle Excellence Award: Technologist of the Year: Cloud Architect by Oracle Magazine. Kai has shared all his technical articles and conference presentations on his Oracle blog: https://kyuoracleblog.wordpress.com/.

Papers

Eliminating database I/O latency with Oracle 21c/19c Persistent Memory features

Event: Connect 2022
Stream: Architecture, Cloud Database & Technology

Database I/O latency often posts significant performance bottlenecks on database systems. Oracle Persistent Memory database and the Oracle Memory Speed File system(OMS) introduced Oracle Database 21c and 19c respectively provide the support for Persistent Memory (PMEM) as database storage and significantly reduce database I/O Latency. By mapping database buffer cache directly onto Persistent Memory storage, Persistent Memory Database directly uses data residing in the Persistent Memory storage and eliminates the latency of copying data into the database buffer cache. This session discusses these two features in detail and how we leverage them to achieve the extremely low latency and high throughputs in a high performance database system. This session provides the step by step how-to details of implementing these two Persistent Memory database features, also provides the use cases & performance benefit studies of leveraging these two features in a high performance database system.

Achieving Extreme Scalability &Total Fault Isolation with Oracle Sharding 21c

Event: Connect 2022
Stream: Architecture, Cloud Database & Technology

Oracle sharding architecture horizontally partitions data across discrete Oracle Databases (shards) that also collectively form a single logical database. In this session we will discuss our experience of using Oracle sharding for the business including: what type of application best fit to Oracle sharding, configuring sharded database with HA replication for massive scalability and complete fault isolation and some lessons learned. We will also discuss some new sharding features in Oracle 21c such as sharding with database in Persistent memory. Furthermore we will discuss our experience with Oracle sharding in Oracle Cloud Infrastructure including docker base deployment of Oracle Shard database and Oracle sharding database on Kubernetes. The audiences will not only learn our experience of using Oracle sharding to achieve massive scalability and complete fault isolation. They will also learn Oracle sharding in Oracle Cloud Infrastructure such as Oracle sharding in docker and Kubernetes platform

Oracle In-database Machine Learning with OML4SQL/OML4Py in Oracle Autonomous Databases

Event: Connect 2022
Stream: Emerging Technologies

As next generation enterprise applications are more data driven and more intelligent, advanced analytics and machine Learning bring great business value. Oracle In-Database machine learning moves the algorithms to database where the data is stored. In this session, we will discuss how Oracle In-database machine learning is provided in Oracle Database including Oracle Autonomous Database with scalability, simplicity and high performance. We will discuss how to build, evaluate and deploys machine learning models natively with large enterprise data using Oracle Machine Learning for SQL(OML4SQL) and Oracle Machine Learning for Python (OML4Py). This session will take some use cases as examples to show the process of machine learning: analyze and prepare data set; build and evaluate and apply machine learning model with Oracle Autonomous Databases environment. Some Oracle 21c new features such as AutoML will also be discussed.