At a glance
Manetu offers a secure data privacy platform that helps consumers and businesses manage personal data.
Its microservices application needed fine-grained trans-action support, secondary indexes, and triplestore capabilities, all while being able to scale to billions of encrypted identities.
With YugabyteDB, Manetu not only achieves global scale and high performance, it also enabled new time series use cases that were not possible with other databases.
Manetu puts consent management in consumers’ hands, enabling consumers to control their data and enterprises to comply with data privacy laws.
New laws in California, Europe and elsewhere require that businesses allow consumers to access the data held about them. Manetu’s Consumer Privacy Management platform includes a consumer-facing portal that gives consumers control and automates fulfillment of these requests; as well as a business-facing portal that simplifies enterprise compliance by tying together all sources of PII into a single, secure, state-of-the-art privacy vault.
The Manetu platform is made up of microservices, written in a mix of Clojure and Go on Kubernetes. Examples of these services include an attribute service, the core of the platform where the PII is encrypted and stored, and an activity log service that provides an activity graph to consumers and administrators. These services are powered by YugabyteDB.
With its current design the Manetu platform is proven at nearly 7 million identities and architected to scale to well beyond 20 million, into billions over time. Manetu is trusted by customers and partners worldwide, including Hedera Hashgraph, Odgers Berndtson, Pickstar, and more.
Challenges
- With their previous Cassandra experience, Manetu knew they needed a Cassandra-like model to support the triplestore for the attribute service, but was concerned Cassandra wouldn’t be able to scale without operational complexity
- Capabilities like fine-grained transaction support and secondary indexes were required for their attribute service, but were not supported in Cassandra
- Their new activity log service required a time series database, but due to the high cardinality of data, neither InfluxDB nor TimescaleDB could scale to meet the ingest rate or required latency to generate the graphs they needed to present in their dashboard
We were not able to use a commercial off-the-shelf time series database like Influx or Timescale. When I tried to put scale on them, the system completely fell apart; it couldn’t keep up with the cardinality of our data. The only solution that was fast enough was to write schemas on top of YugabyteDB, and it works really well.
We’re not aware of another database that would allow us to scale to every person in the world. We have global aspirations and YugabyteDB as part of our architecture could see us well down the road into billions of identities and managing events for all of them.
Greg HaskinsChief Technology Officer,
Manetu
Conor AllenHead of Product,
Manetu
Yugabyte Solution
96 total cores &
24 TB of data
3 node clusters in
single region, multi-
zone deployments
Yugabyte + Kafka
as primary
persistent layer
Easily deployed
and managed in
Manetu’s AWS VPC
Technical Results
- Easy to scale out with
transactions and secondary
indexes - Effectively manages time
series workloads with
high-cardinality datasets - High performance reads and writes in a microservices, CQRS architecture
- Able to ingest high volumes of data quickly during onboarding events (bringing on new customers and the identity data they hold)
Business Results
- Comprehensive multi-region deployment options
to enable compliance across the globe as
business grows - Flexibility to support multiple
use cases, including sensitive PII
and time series data - Responsive technical
support - Future-proofing business for
billions of identities
Manetu has built a custom RDF triple-store featuring transparent cryptography, individual keys for every relationship, and de-identification features to serve our customer’s privacy needs from the cloud. Serving this type of data securely and at scale while maintaining SLAs is non-trivial. However, YugabyteDB’s high performance, compatibility with the larger CQL ecosystem, and inherent ease for maintaining highly available multi-AZ deployments and ‘push button’ elastic scale made it an easy choice to underpin our offering, and allowed us to build a world-class service. It also helped that they have a program for startups!
Greg HaskinsChief Technology Officer, Manetu
See more from Manetu
Developing Cloud-Native Spring Microservices with Distributed SQL, Cluster-Aware JDBC Drivers, R2DBC and Istio
In this presentation, Nikhil Chandrappa from Yugabyte’s Ecosystem Engineering team, will show you by example how Spring can be used in the development of highly scalable and elastic microservices. By examining the code and architecture of real world examples, he’ll show you how PostgreSQL-compatible distributed SQL backends, cluster-aware JDBC drivers, reactive programming with R2DBC and Istio can be combined to deliver these services rapidly.