Stateful HTAP (Hybrid Transactional and Analytical Platform)

Cyoda was built to fulfil the technical requirements to implement virtually any organisation’s data processing, batch or real-time and analytics/reporting, with transactional consistency, but without restrictions. A horizontally and elastically scalable platform without limitations to cut the total cost of ownership, implementation, running and change.

What is it not?
  • It’s not built on or with Hadoop.
  • It’s not an appropriate technology for Ultra-low latency operations such as high-frequency trading.
Distributed, scalable, reliable
  • No single point of failure, no bottlenecks
  • Dynamically scalable to meet fluctuating demands
Consistency
  • ACID consistency through distributed transactions
  • Processing dependency
Event-based
  • Distributed, scalable
  • Maintains sequence at the object level
Stateful
  • Write only store
  • Object persistence, read and write is taken care of so you model only at the object level.
Data Processing
  • STP in user-defined event-driven BPM (Business Process manager) workflow via state-machine (Rules Engine)
  • Real-Time processing
  • Scales linearly to allow unlimited volumes of data to be processed
  • Workflow can be defined by a non-technical user.
  • Transparency of the system
  • Lower cost of change
Schema agnostic, Schema on read
  • No requirement for unified data model
  • Leave data in their original format
  • Reduce data duplication
  • Reduce implementation time
  • Model data combining different complex data models and version how and when you want.
Queries/reports/aggregations
  • OQL (Object Query for defining) for complex queries and aggregations
  • Distributed report load for linear scalability to run reports as fast as required
  • Transactional consistency for accurate ad-hoc reporting on live changing data
  • As-at and bi-temporal reporting.
  • Composite range indexing for constant time data extraction regardless of volumes stored (not brute force reporting)
  • Can index raw data as well as secondary generated data
  • Enables fast, ad-hoc, flexible reporting, both current and as-at, against arbitrarily complex data and queries. OQL implementation (Object Query Language)
  • Combine multiple complex data models and version into a single combined report, define the query and aggregation on the virtual model.
More Info

Further reading:

Overview

STP Processing

Easy Integration

Distributed Queries

Scale Without Limits

Data Consistency