Reading Notes: Lamport Clocks and Happens-Before
Notes on Lamport's 1978 paper introducing logical clocks and the happens-before relation, with observations on how the idea maps onto modern distributed databases.
Tag
#Distributed Systems
8 entries
Adaptive Consensus Protocols Under Partial Synchrony
We present AdaptRaft, a variant of the Raft consensus algorithm that dynamically adjusts its timing parameters using a lightweight online learning model, reducing tail latency by up to 34% in geo-distributed deployments.
AdaptRaft
A modified Raft implementation with an online bandit learner that adapts election and heartbeat timeouts to observed network conditions. Companion codebase for the SOSP 2024 paper.
Working Notes: What Raft Actually Guarantees
Informal notes clarifying which safety and liveness properties Raft provides under which assumptions, written while preparing a talk on adaptive consensus.
Byzantine Fault Tolerance in Practice: A Survey of Modern BFT Protocols
A structured survey of Byzantine fault-tolerant consensus protocols deployed since 2015, identifying the practical trade-offs between safety, liveness, and throughput that are often obscured in theoretical treatments.
Distributed Systems — Hilary Term 2024
Graduate teaching assistant for the Distributed Systems course. Eight tutorials covering consensus, replication, fault tolerance, and consistency models.
DistBench
A configurable benchmarking harness for distributed consensus protocols. Supports pluggable protocol implementations, synthetic and replay workloads, and structured JSON output for reproducible experiments.
Learned Indices for Distributed Key-Value Stores
We explore the applicability of learned index structures to distributed key-value stores, demonstrating that model-based indices can reduce lookup latency by 22% with a modest increase in memory footprint on skewed workloads.