The WSO2 Research Team are Unsung Heroes of Performance
Our customers run billions of transactions on or through our software. Marginal improvements can have an outsized impact on the total cost customers must pay to operate our software. For this reason, it’s almost always the case that stress and performance analysis are part of the evaluation cycle of most of our customers.
It is not widely known, but WSO2 has a dedicated research team. Overseen by Srinath Perera and Malith Jayasinghe, the research team fills a valuable and essential role in our performance engineering efforts. This effort requires a lot of attention and the team 3 senior staff members and 4 research engineers, who are recent graduates from Sri Lanka universities achieving the highest standards of attainment including valedictorians.
Their primary focus is performance engineering and acting as the governing body for WSO2 product performance. This involves:
- Own platform performance by coordinating, mentoring and providing expertise on performance-related activities (performance testing, performance optimization, troubleshooting, automation and capacity planning).
- Provide assistance in resolving complex performance-related support
- Research on the performance behaviors of our products, theoretically as well as experimentally and develop new ideas to improve performance. The following are some contributions in 2018.
In 2018, the team helped resolve 34 complex performance-related support issues, developed 2 papers on performance behavior currently under review for publication, and supported two thesis projects.
Additionally, this team was essential in the design and construction of geo-velocity adaptive authentication in our Identity Server, Bayesian learning algorithms for streaming machine learning, and incremental checkpointing for Siddhi.
They take this work and evangelize to developer audiences having published 20 articles, collaborate with university interns on various project, and supported 6 masters / PhD theses.
Are you ready to be inspired? Then you should catch up on the amazing achievements and writings made this year by this team.
Talks
- Keynote: Engineering High Performing Systems, 11th International Conference 2018 (KDU-IRD 2018),
- Privacy in Bigdata Era, National IT Conference 2018, https://www.slideshare.net/hemapani/privacy-in-bigdata-era
- Blockchain, Impact, Challenges, and Risks, Launch of MSc in Cyber Security and Forensics, Informatics Institute of Technology Sri Lanka, https://www.slideshare.net/hemapani/blockchain-impact-challenges-and-risks
- “Java in Flames”, Java Colombo Meetup
Research Publications
- Ballerina: A Next Generation of Flow Languages
- Recent Advancements in Event Processing. ACM Comput. Surv. 51, 2, Article 33 (February 2018), 36 pages. DOI: https://doi.org/10.1145/3170432
- (Under Submission) Understanding the Workload Characteristics of Server-Based Workloads Under Different Heap Sizes, Venue: ISPASS 2019
- (Under Submission) Group Wallets: Hierarchical Deterministic Wallets for Anonymous Group Transactions, Venue: FC2018
- (Under Submission) Length Pre-Image Attack for SHA-2, Venue: WISTP2018
- (Under Submission) MCMC methods and Variational Inference for Bayesian Regression Models, Venue: AISTATS 2019
- (Under Submission) Streaming Stochastic Variational Bayes; An Improved Technique for Bayesian Inference with Data Streams, Venue: AISTATS 2019
- (Under Submission) Low Latency Technique for Handling Out-of-Order and its Application for Event Pattern and Sequence Detection Queries of Data Stream Processing, Venue: EDBT2019
- (Under Submission) An Improved Adaptive Authentication Mechanism Based on Geovelocity, Venue: ICC2019
Articles: Infoq, Infoworld, Datanami, Programmableweb
- Infoworld: It’s time we tapped APIs for business analytics
- Infoq: How to Choose a Stream Processor for Your App
- Datanami: Five Ways to Apply Streaming Analytics Now
- Datanami:Streaming SQL for Real-Time Analytics
- ProgrammableWeb: How APIs Can Drive The Adoption of Meaningful Business Analytics
DZONE articles
- Apache Storm vs WSO2 Stream Processor, Part 1
- Apache Storm vs WSO2 Stream Processor, Part 2
- Is Your Stream Processor Obese?
- Bayesian Learning for Machine Learning part 1 — Introduction to Bayesian Learning
- Performance Improvements via JIT Optimization
- Improving the Performance of WSO2 MB by Controlling Buffer Limits
- Moving Towards Self-Adaptive Systems: ML-Based Auto-Tuning in Middleware
- The Rise of Self-Sovereign Identity: Hyperledger Indy
- Ballerina Services in Serverless World
KDnuggets articles
WSO2 Articles
- Distributed Stream Processing with WSO2 Stream Processor
- The Impact of Meltdown, Spectre and JDK Patches on the Performance of WSO2 API Manager
Medium Stories
- A Gentle Introduction to Stream Processing, (9K views, 370 calps, #2 at Google for “Steam Processing”)
- Hackernoon: 30 Shared Principles for discussing Software Architectures (6K views, 600 claps)
- Draconian, free, or nanny state: Concurrency ideologies in Java, C#, C, C++, Go, and Rust, (12K views, 718 claps)
- Linux Performance Observability Tools (6.4K views, 252 claps)
- Java Flight Recorder Cheat Sheet (2.1K views, 56 claps)
- Stack Depth in Java Flight Recorder (198 views, 0 claps)
- Performance Optimization via JIT Optimization (1.1K views, 148 claps)
- Moving towards self-adaptive systems (120 views, 151 claps)
- How to create a RPC enabled private network in Ethereum (1.8k views, 76 claps)