Skip to main content

EloqData on Monster Scale Summit 2025

· 4 min read
EloqData

Redis is widely used for caching and real-time data processing, but its scalability limitations create significant challenges for modern applications. At Monster Scale Summit 2025, we introduced EloqKV, a high-performance, distributed transactional database that overcomes Redis' scaling bottlenecks by leveraging ScyllaDB. This blog post explores how EloqKV enhances scalability, performance, and transaction capabilities while maintaining Redis API compatibility.

Overcoming Redis Scaling Bottlenecks with EloqKV on ScyllaDB

Introduction​

At Monster Scale Summit 2025, Eloqdata is honored to be the featured speaker, presenting how EloqKV on ScyllaDB, our high-performance, distributed transactional database, tackles Redis scaling bottlenecks. Hosted by ScyllaDB, a distributed low-latency NoSQL database, the summit showcases cutting-edge solutions for extreme-scale engineering.

Watch the video.

Visit the official Monster Scale Summit website.

EloqKV vs Redis Transaction

The Redis Scaling Challenge​

Redis is a popular in-memory database, but scaling it presents significant issues:

  • Scale-Out Complexity: Redis clustering is full of pitfalls, including its lack of support for cross-node operations and the increasing overhead of its gossip protocol as the cluster size grows.
  • Data Volume Limitations: Redis is purely in-memory, making it expensive and impractical for large datasets.
  • Single-Threaded Execution: Redis handles requests in a single thread, limiting its ability to scale up efficiently.
  • Transaction Limitations: Redis does not provide true ACID transactions, making complex workflows difficult.
  • Cache Coherence Issues: Redis and its underlying store often suffer from consistency problems, often necessitating extra application-level logic to ensure data integrity and accuracy.

These issues force teams to make trade-offs between cost, performance, and ease of use. EloqKV on ScyllaDB is designed to remove those trade-offs by combining Redis' API with scalability and transaction.

How EloqKV on ScyllaDB Solves Redis Bottlenecks​

EloqKV is a full-featured, distributed ACID transactional database that is compatible with the Redis API. By integrating with ScyllaDB, it delivers:

1. Multi-Tier Storage for Cost Efficiency​

  • Hot Data: Cached in EloqKV’s buffer pool for fast access (~100us).
  • Cold Data (Option I): Stored in Key-Value engines like ScyllaDB, ensuring high-performance cache miss reads.
  • Cold Data (Option II): Offloaded to object storage, reducing total cost of owership.

2. High Throughput with Thread-Per-Core Execution​

EloqKV follows a multi-threaded, thread-per-core model akin to ScyllaDB, though it is implemented independently without relying on Seastar, ensuring:

  • Full CPU utilization for maximum performance.
  • 10x faster than Redis, and delivers performance comparable to DragonflyDB, while providing transactional capabilities and native scale out features.

3. True Distributed Transactions​

Unlike Redis, EloqKV supports:

  • Durable Cross-AZ Write-Ahead Logging (WAL).
  • Popular Isolation Levels/Concurrency Control Protocols.
  • Cross-node Multi/Exec & Lua Scripts.
  • Explicit Begin/Commit/Rollback transactions (Unique).

4. Seamless Elastic Scalability​

  • Elastic scaling across all resource types: CPU, memory, storage, and logs.
  • Optimized component scaling tailored to your workload for reduced Total Cost of Ownership (TCO).
  • Easy integration with cloud-native infrastructures.

Beyond Redis: The Future of EloqData​

EloqKV is more than just a key-value store. At EloqData, we invent a new modular database approach, enabling different compute models to share the same data substrate. This vision extends beyond a single API, with:

  • EloqSQL: MySQL-compatible distributed RDBMS.
  • EloqDoc: MongoDB-compatible document store.
  • EloqKV: Redis-compatible key-value store with transactional support.

By leveraging open-source execution engines, we ensure compatibility with existing ecosystems while avoiding redundant reinvention of wheels.

Conclusion​

Redis has been a go-to for in-memory caching, but its scaling challenges limit its long-term viability for demanding workloads. EloqKV on ScyllaDB offers an alternative that provides high performance, ACID transactions, and cost-effective scalability—without the trade-offs.

If you’re looking to break free from Redis scaling constraints, EloqKV is the future. Let’s build the next-generation database together and stay connected: 🔗 GitHub