Combined power of Kafka and Flink
These underlying technology choices are at the heart of Digazu’s non-functional excellence: real-time, low latency data processing at scale.
Kafka is managing the data collection and distribution aspects (stream processing) while Flink focuses on the manipulation of data in streams (data transformation).
Flink is particularly efficient at managing stateful computations which are a must when you want to combine several high-volume sources in real-time and make aggregations typically used for real-time dashboards for IoT.
Both frameworks have extensive communities and ecosystems and are there to stay in the foreseeable future.
Digazu can be deployed on any Kubernetes cluster which are natively supported by all the main cloud providers. Digazu does not lock you in with any cloud provider as the streaming of data it provides can effectively be sourced from and streamed to any environment.
From a deployment perspective, it is very easy to start using Digazu for new projects without impacting any of the existing data infrastructure. Further rationalisations of the environment can be done at any subsequent stage.
The reconciliation of project-level needs with enterprise-level considerations of reuse, access management and controls is an age-old problem in data management. Digazu solves this elegantly with its decentralized approach and full data lineage.
Digazu can also be integrated with existing enterprise resources: monitoring and alerting, meta-data management, authentication, code version control, and API management systems and feeds naturally into existing consumption environments (BI and data science).