Case Study: Docudesk achieves low-latency, high-throughput document management and 99.999% uptime with MySQL Cluster (MySQL)

MySQL Cluster Powers Leading Document Management Web Service

Docudesk, a Dallas-based PDF software developer, launched DocQ to provide a paperless document management and transaction service that preserves PDF quality, security and legal e-signing while replacing fax and courier workflows. The team needed a database that could handle heavy write throughput (about half of workload), deliver millisecond-level latency, meet 99.999% availability, scale linearly, and minimize operational complexity.

Docudesk chose MySQL Cluster over a sharded MySQL + memcached design and deployed it on Amazon EC2/S3. The cluster (2 data nodes, 2 management nodes, 6 SQL nodes) stores metadata, text pages (BLOBs), sessions, ACLs, job queues and billing actions, handling roughly 500,000–1,000,000 queries per day across in-memory and disk tables (about 4 GB and growing). The integrated solution delivered the required low latency, high write throughput and continuous availability with simpler operations, and Docudesk is upgrading to newer Cluster releases and testing online scaling to grow capacity without downtime.

Docudesk

Casey Brown

DBA and Development Manager