Publications

Domestic Non-Peer-Reviewed Papers

• Yoshiaki Ueda, Toshinobu Usui, Construction of a Low-Latency Distributed KVS for Personalized Content Delivery
  • Proceedings of the 82nd National Convention of IPSJ, 2020-02-20, pp. 87-88
  • PDF / Bibliographic Info
• Yoshiaki Ueda, Proposal for Improving DNS Cache Hit Rate and Response Speed for Distributed Systems
  • Proceedings of the 84th National Convention of IPSJ, 2022-03-05, Preprint, Student Encouragement Award

Technical Books (Doujinshi)

• Learning by Building: Raft-based Redis-compatible KVS in Go
  • TechBookFest 16 (Technical Doujinshi / 79 pages, ver 1.2)
  • Goal
    • An introductory hands-on book that teaches distributed systems, consistency models, and Raft by implementing a Redis-compatible KVS.
  • Structure
    • The first half covers distributed systems fundamentals, Redis/Redis Cluster strengths and limitations, consistency models, and Raft core concepts (leader election, log replication, safety).
    • The second half connects theory to implementation through a full Redis-compatible KVS build.
  • Hands-on implementation (Go)
    • Uses hashicorp/raft, implementing the datastore, Snapshot, and StateMachine (Apply/Snapshot/Restore).
    • Implements RESP with redcon and supports GET/SET/DEL.
    • Handles follower-side requests via MOVED redirection to the leader, validated in a 3-node setup.
  • Production context
    • Links the design choices to real strong-consistency/high-availability use cases, including PayPay’s TiDB adoption.
  • TechBookFest Product Page / BOOTH

Presentations

• How Cloud-Native Databases Overcame Their Constraints: Scalable Architecture Design Principles Through Their Evolution

  • Cloud Native Kaigi 2026 (2026-05) / Session Page
  • Traced the evolution of sharding, compute-storage separation, distributed consensus, and global ordering to explain how cloud-native databases address the constraints of write scaling, I/O, and strong consistency.

• NewSQL: Scalable Architecture with Storage Disaggregation and Distributed Consensus

  • Database Engineering Meetup #9: NewSQL (2026-03) / Event Page
  • Explained the designs behind NewSQL scalability and their operational considerations through compute-storage separation, storage-side WAL application, Raft, Multi-Raft, leader leases, and hybrid logical clocks.

• The Mechanism Behind NewSQL and Distributed Databases: Raft

  • DB Tech Showcase 2025 (2025-07)
  • Explained the internal structure of the Raft algorithm and its application in NewSQL (CockroachDB/TiDB, etc.) from an application engineer’s perspective at one of Japan’s largest database conferences.
    • Related paper: In Search of an Understandable Consensus Algorithm (Extended Version)

• Designing an Event Bus for Loosely Coupled Schema-Driven Development

  • Platform Engineering Kaigi 2025 / Session Page
  • Presented a case study of migrating from monolith to microservices at hacomono. Covered dependency elimination using the Fire-and-Forget pattern, schema management with Protocol Buffers, and governance via compatibility checks in CI.

• What is Raft? Understanding its Strengths and Weaknesses from the Mechanism

  • Visually explained the behavior of the distributed consensus algorithm Raft (leader election, log replication) and the trade-offs between consistency and availability (3,600+ views).
    • Related paper: In Search of an Understandable Consensus Algorithm (Extended Version)

• Why Did Spanner Need an Atomic Clock?

  • Explanation of the physical and logical necessity of the TrueTime API and atomic clocks supporting Google Spanner’s External Consistency.
    • Related paper: Spanner: Google’s Globally-Distributed Database

• Reading the Paper on AWS’s Durable Redis-Compatible KVS: MemoryDB

  • Explanation of the persistence log mechanism of in-memory DBs and architectures for achieving high availability.
    • Original paper: Amazon MemoryDB: A fast and durable memory-first cloud database

• Building a Low-Latency Distributed KVS for Personalized Content Delivery (VRChat Ver)

  • Explanation of optimization approaches for personalized content delivery.
    • Related explanation of my own paper: Construction of a Low-Latency Distributed KVS for Personalized Content Delivery

• Understanding Challenges and Insights Moving from Aurora Serverless to v2 from the Paper

  • Explains Aurora Serverless scaling issues and the improvements that led to Aurora Serverless v2.
    • Original paper: Resource management in Aurora Serverless

• Reading the Paper on AdaptSize Supporting Akamai’s Cache Efficiency

  • Explanation of the theory and practice of the AdaptSize algorithm for optimizing CDN cache efficiency.
    • Original paper: AdaptSize: Orchestrating the Hot Object Memory Cache in a Content Delivery Network