Capabilities for Operational/Transactional Workloads

• Deep Integration with HBase for row-wise access
• Leverages and optimizes to strengths of underlying engine
• Truly Distributed Transaction Manager
• Provides multi row, table, and statement ACID transactional support
• No bottlenecks to scaling in volume, concurrency, and size of transactions
• Very efficient implementation, even matching HBase on single row transactions
• Multi-Dimensional Access Method
• Efficiently uses multi-column clustering index, even without predicates on leading columns, to retrieve just the rows needed and avoid scans on very large datasets
• Substantially reduces need for indexes
• Facilitate salting and computed (e.g. divisioning) columns as clustering key prefixes
• Salting facilitates partitioned parallelism across the cluster, with ease of rebalancing as cluster is expanded
• Time series data clustered using divisioning, for efficient access

Capabilities for BI and Analytical Workloads

• Deep Integration with Apache ORC and Parquet for column-wise access
• Leverages and optimizes to strengths of underlying engine
• Sophisticated and mature SQL Optimizer, that learns from previous executions, to generate optimal query plans
• Massively parallel data flow run-time engine for fast response times, while optimizing resource utilization, overflowing to disk automatically when there is memory pressure
• Adaptive Degrees of Parallelism
• Uses only the number of cores needed to execute a query, instead of using the entire cluster, achieving very high levels of concurrency, resiliency, and resource efficiency
• Skew Buster
• Eliminates skew in joins, rampant at higher scales, dramatically reducing response times and increasing the efficient and balanced use of the entire cluster

Hybrid Transactional/Analytical Processing Capabilities

• Optimizes for operational and analytical workloads; depending on the query chooses
• Serial or parallel plans
• Appropriate join type: nested, probe cache for nested, merge, matching partition, repartitioned hash, replication by broadcast hash, inner / outer child broadcast
• Uses risk premiums for nested and merge joins, and serial plans to avoid sub-optimal plans
• Hot and Cold Data support
• Expression based copying / moving of data from HBase to Apache ORC / Parquet
• Access to HBase and / or Apache ORC / Parquet depending on type and focus of query
• Multi-Tenancy and Workload Management System (WMS)
• Allows allocation of appropriate cluster resources to operational and analytical workloads based on desired priorities and service levels desired
• WMS gathers detailed workload metrics and facilitates threshold based alerts / actions
• Security administration by tenant

Capabilities common to all workloads

• Advanced Table Mapping User Defined Functions
• Provides extensibility for integration for
• Streaming – such as with Kafka
• No SQL data model integration – SOLR for search, JanusGraph for graph
• Analytical capabilities – such as Spark ML, R, TensorFlow
• Other databases, using JDBC – e.g. Oracle
• Comprehensive ANSI SQL and common Oracle, Teradata, other database function support, including support for ANSI OLAP functions
• Java Stored Procedures and C++ and Java User Defined Functions / Table Mapping UDFs
• ANSI SQL GRANT/REVOKE at user or role level from schema to object level, as well as for operational capabilities
• Connectivity services for JDBC, ODBC, ADO.NET, and REST API clients, providing scale, load balancing, high availability, and connection pooling
• Integration with Hibernate to provide ORM application development support
• Loading tool, as well as bulk and incremental load capabilities, even on indexed tables

Enterprise Capabilities

• On premise or cloud deployment
• Provides disaster recovery with zero lost transactions, and scaling of reads and writes across multiple clusters and data centers
• Active-active multi-master distributed synchronous replication at the table level
• Provides point-in-time recovery to recover from operational errors leveraging transactionally consistent snapshots, transactional logs, and snapshots after non-transactional bulk loads
• Table level backup and restore
• Scales elastically, enabling dynamic online expansion or reduction of nodes or storage
• Supports the separation of compute and storage to scale them independently
• Enterprise security integrating with Apache Sentry, Secure Hadoop to leverage Kerberos for authentication, multi-domain Active Directory or LDAP; Encryption functions
• Database Designer and Manager tools for developing, monitoring, and managing databases

EsgynDB is ideally suited to supporting Hybrid Transactional and Analytical Processing (HTAP) or Translytics or Hybrid Operational-Analytical Processing (HOAP) - terms coined by Gartner, Forrester, and 451 Research respectively, for the same concept of leveraging one database platform to handle multiple mixed workloads concurrently ranging from short-running transactional or operational queries to long-running or complex analytic queries at the same time.  EsgynDB can handle this with support for massive concurrency demands, and on very large datasets. 

EsgynDB scales out linearly to handle big data, and provides enterprise class features whether that be SQL maturity, disaster recovery, or security features, with the ease and reliability of a traditional SQL RDBMS but with the scalability and performance of NoSQL engines.

Ideally suited to scenarios where high speed continuous data ingestion is required while in parallel doing real-time reporting and alerting, coupled with deep analytics.  Many IoT use-cases fit this model.

EsgynDB is also very well suited to modernizing data lakes, by bringing the power of a fully featured SQL engine to where the data resides, and being able to query across multiple disparate storage engines or storage formats.

EsgynDB's customers are worldwide, with several in the Global Fortune 500, including one ranked in the top 5.  Use cases span various verticals from Financial Services, to Telecom, to Security, to Power and Energy Companies and more.

Customers that can be publicly referenced are listed at www.esgyn.com

EsgynDB Enterprise Edition is the commercially supported version of Apache Trafodion with some additional manageability features, with both 24x7 and 8x5 support models available.

EsgynDB Enterprise Advanced Edition contains additional value-added features needed for more complex or demanding scenarios, such as multi-datacenter support or multi-tenancy support.

EsgynDB is available on premise or in the cloud on a per-node per year license and support subscription model.