As Whitepages Expands, It Relies on Redis on Flash to Keep Operational Costs Low and Performance High

• 分析与建模 - 预测分析
• 基础设施即服务 (IaaS) - 虚拟私有云
• 平台即服务 (PaaS) - 数据管理平台

• Professional Service
• Software

• 商业运营
• 质量保证

• 预测性维护
• 实时定位系统 (RTLS)
• 远程资产管理

• 云规划/设计/实施服务
• 软件设计与工程服务
• 系统集成

Users of Whitepages’ people search and identity verification services expect instant results from a single piece of contact information, which requires a highly efficient and fast database system. The company's proprietary Identity Graph™ database, which houses billions of person-identity records, generates a high volume of calls to Redis, especially during peak times. Initially, Whitepages used Amazon ElastiCache to store IDs, but as the company expanded globally, this solution became resource-intensive and unwieldy. The company evaluated MongoDB, Apache Cassandra, and Couchbase as potential replacements, but none could handle the full dataset size while providing the required single-digit latency. Whitepages' applications are extremely latency-sensitive, and customer SLAs promise sub-hundred millisecond latency, leaving no room for error.

Whitepages is a global leader in digital identity verification services for businesses and consumers. The company offers innovative people search tools powered by more than five billion global identity records. As Whitepages expands its identity dataset beyond North America, it increasingly relies on advanced database solutions to maintain high performance and low operational costs. The company's proprietary Identity Graph™ database is a key component of its services, enabling comprehensive and verified views of individuals' public footprints, including past addresses, family members, and legal documents. Whitepages' commitment to customer experience and performance drives its continuous search for efficient and scalable database solutions.

After several failed attempts with other NoSQL databases, Whitepages began researching enterprise-grade Redis from Redis Labs. Redis on Flash, a unique offering from Redis Labs, caught their attention. Redis on Flash stores hot values and keys in RAM and cold values in cost-effective flash-based SSDs, drastically reducing operational costs without compromising performance. Whitepages now uses only 30% of the RAM storage it previously used, resulting in significant infrastructure savings. The company stores more than three terabytes of data in Redis on Flash, deployed in its virtual private network. Replication is achieved through a master/slave configuration across two availability zones, ensuring high availability. Redis on Flash provides the single-digit latency required to meet customer SLAs, allowing Whitepages to deliver top-of-the-line services without compromising performance. The company is also exploring additional features offered by Redis Labs, such as Redisearch and geo-replication, to support future plans.

• Whitepages achieved significant operational savings by using Redis on Flash, which stores cold values in flash-based SSDs, reducing the need for expensive RAM.
• The company maintained high performance with Redis on Flash, even for datasets measured in terabytes, ensuring single-digit latency and meeting customer SLAs.
• Replication through a master/slave configuration across two availability zones ensured high availability and reliability of the database.

• Whitepages now uses only 30% of the RAM storage it previously used, resulting in hundreds of thousands of dollars in infrastructure savings each year.
• The company handles upwards of 200,000 calls to Redis per second during peak volumes, maintaining single-digit latency.