On Supporting Digital Identity at Whitepages. Q&A with Priya Balakrishnan
Q1. What is the business of Whitepages?
Whitepages is a global leader in digital identity verification services for businesses and consumers. The company’s innovative people search tools are powered by more than five billion—and growing—global identity records. Users of Whitepages services are able to instantly access a comprehensive and verified view of a person’s public footprint with a simple search. In addition to that, Whitepages also provides a service to organizations that connects to their API for in-depth identity verification.
Q2. Why are Whitepages’ applications extremely latency sensitive?
When organizations leverage Whitepages API for identity verification, it is one of the many calls that’s made in the back end for that particular transaction. With SLA’s promising sub-hundred millisecond latency, they needed a database platform that will dramatically minimize the end-to-end latency with sub-millisecond response times.
Q3. Whitepages has chosen Redis Labs Enterprise database solution to power the company’s proprietary Identity Graph. What is Whitepages` Identity Graph useful for?
Whitepages identity graph houses billions of person-identity records with linkages that connect people, phones, addresses, emails, and IPs to help businesses confidently assess and verify consumer identities. With Redis Enterprise and Redis on Flash, Whitepages has achieved over 70% of infrastructure cost savings.
Q4. Redis provides sub-millisecond latency. How?
Redis Enterprise delivers exceptional linearly scaling high performance with its shared nothing cluster architecture, automatic re-sharding and rebalancing to utilize your infrastructure to the best possible extent. It also includes enhanced connection management and request scheduling and high-performance pipeline execution and cluster support for every client. Most recently, Redis Enterprise has been benchmarked to demonstrate true linear scaling – going from 10M ops/sec with 6 AWS EC2 instance to 30M ops/sec with 18 AWS EC2 instances to 50M ops/sec with 26 AWS EC2 instances.
Additionally, Redis Enterprise extends RAM onto Flash SSDs while delivering the same sub-millisecond latency that Redis is well known for. This technology offers the cost-effectiveness of Flash, with the blazing fast low latencies of Redis in RAM, delivering up to 80% on infrastructure costs with large datasets.
Q5. How is it implemented in the use case of Whitepages?
Whitepages uses Redis on Flash to intelligently tier their data storing frequently used hot values and keys in RAM and cold values on Flash SSDs. This drastically reduces their overall utilization of RAM to only 30% of what was previously used, dramatically saving them hundreds of thousands of dollars in infrastructure savings each year.
Q6. How do you plan to handle 150,000 to 200,000 calls per second with a three-terabyte database?
A single name query into the Whitepages API gives rise to many additional calls into their connected identity graph database. During peak volumes, this can translate to upwards of 200,000 calls to Redis per second. Redis Enterprise can seamlessly scale up by simply adding shards to your database without adding nodes to your cluster and scale out by adding node(s) to the Redis Enterprise cluster, rebalancing and resharding. In addition, Whitepages achieves replication through a master/slave configuration across two availability zones and reads go to both to ensure high availability. This allows Whitepages to handle large throughputs while delivering sub-millisecond latency on reduced infrastructure footprint.
Q7. How do you plan to reduce Whitepages’ operation costs without sacrificing performance?
Redis on Flash extends Redis from RAM to a combination of RAM and Flash SSDs, using an intelligent approach to tiering, delivering stable performance and delivering sub-millisecond latencies. With Redis on Flash, Whitepages has gained enormous infrastructure cost savings compared to any other NoSQL databases, including MongoDB, Apache Cassandra, Couchbase and AWS ElastiCache.
Q8. What other use cases do you have for Redis on Flash (RoF)?
Redis on Flash delivers significant cost savings by extending RAM onto Flash SSDs and intelligently and automatically tiering data based on frequency of access, without any compromise to performance. Even when running Redis in RAM, additional savings can be incurred with Redis Enterprise because it enables multiple Redis instances to be run on the same server, utilizing all the cores more efficiently than single-threaded Redis can.
Qx. Anything else you wish to add?
Redis Enterprise is uniquely suited to develop modern applications faster with many built-in capabilities for user session management, high speed transactions, fast data ingest, content caching, embedded analytics, search, time series, graph, machine learning model serving and many more. Redis Enterprise has thousands of customers in every industry and serves their needs in a range of solutions from e-commerce, Personalization, IoT, Metering, Social Apps, Fraud Mitigation and many more. Redis Enterprise is available in any deployment model of choice, on any public cloud, as a fully managed database-as-a-service inside VPCs and as downloadable software.
Priya Balakrishnan, Sr. Dr. Product Marketing at Redis Labs.
Sponsored by Redis Labs