Esgyn

Last Updated: 4th August 2020
Analyst Coverage: Philip Howard

Esgyn Corporation, the developer of EsgynDB, is based in the United States and China. The company was founded by ex-HP staff that had worked on successive generations of that company’s database technology, starting with Tandem NonStop SQL, via HP’s enterprise data warehouses, to EsgynDB, which represents the 4th generation of development, and which is a SQL on Hadoop engine focused on supporting hybrid environments where you want to combine transactional and analytic processing with mixed workloads and multi-structured datasets, though it can and does support either of these as independent requirements. It is a commercial implementation of the Apache Trafodion database, to which Esgyn is the prime contributor.

The product is available both in the cloud (and as a part of Esgyn Strato, which is a cloud-based platform offering available as a managed service) and on premises. It is also, in conjunction with Supermicro, available as an appliance. EsgynDB has been deployed in a variety of areas, including Internet of Things (IoT), Banking and Insurance, Telecom, Manufacturing, Internet Security and Smart Cities. Existing clients include a major US bank, China Telecom, China Mobile, Foxconn and Webroot, amongst others.

Company Info

Headquarters: 691 South Milpitas Boulevard, Suite 117, Milpitas, CA 95035
Telephone: +1 408 800 6916

EsgynDB

Last Updated: 24th June 2019

What is it?

Fig 01 The history of ESGYN

EsgynDB is a massively parallel processing (MPP) based database that typically leverages Trafodion tables, but which also supports Hive, HBase (either natively or as the underpinning for Trafodion), Parquet, ORC and Avro formats. The product supports extensible user defined functions (UDFs) that allow external environments to be treated as tables. Currently supported external environments include Apache Solr and JanusGraph (which is a Linux Foundation open source graph database). In the latter case, you can embed Gremlin code into your UDF so that you can join graph data with data managed directly by EsgynDB. The product also integrates with and leverages technologies such as Hibernate, Spark (including Spark MLlib), TensorFlow (which can be pushed directly into the engine via the Table Mapping UDF), R (ditto) and various streaming engines (including Kafka with parallel streaming). In terms of other features, the product supports a wealth of capabilities, including secondary indexes, the ability to assign parallel resources as required (a part of its workload management as well as the database optimiser), the sorts of high availability that you would expect from a product with a Tandem heritage (with both single and multi-master synchronous replication), load balancing and so on. There is a distributed query manager to support transaction processing and the product is fully ACID compliant. SQL support includes extensions that support Oracle PL/SQL and Teradata functions, to encourage users to move off those platforms and onto EsgynDB.

Customer Quotes

“Use of Esgyn technology is not only beneficial for making decisions based on business intelligence, but also for saving energy, reducing consumption and leading low-carbon lifestyle.”
Wanda

“EsgynDB accelerated our time to bring new solutions to market due to simplicity of a standard ANSI SQL interface. With EsgynDB we now have the cloud-based architecture we wanted to deploy new and additional workloads quickly and easily.”
Webroot

How does it work?

Fig 02 The architecture of ESGYN

The architecture of EsgynDB is illustrated in Figure 2. Starting at the bottom, storage is based on HDFS and can leverage Trafodion, native HBase or Hive tables. Trafodion and HBase tables are typically used for “hot” data, whilst “warm” data can be stored in columnar format in ORC or Parquet tables. “Cold” data will typically be stored in Amazon S3 or an Azure Data Lake. In other words, the database can act to support multi-temperate data. To support improved performance when row-based data has to be queried, EsgynDB supports a multi-dimensional access method, along with salt keys, that provides a superset of what other vendors refer to as “row skipping” and which improves performance when you need to do whole table scans. The company also provides patented skew detection and prevention to ensure that data is distributed uniformly for joins, which will also help to improve performance and ensure that it is consistent.

Moving up the stack, one thing that may not be clear from this diagram is that “ESP” stands for executor server process. There can be as many of these as the workload needs and they are a key component in enabling parallelism, which is available in multiple forms: pipeline, operator and partition parallelism are all provided. The distributed transaction management is implemented on each node in your cluster, while in-memory processing is used wherever this is feasible, dynamic spill-to-disk capabilities are provided for occasions when memory capacity is insufficient: this enables higher concurrency and improved throughput for high concurrency workloads.

EsgynDB separates storage from compute and one of the consequences of this is that you can have different instances of EsgynDB within the same environment, dedicated to different tasks. Thus, you could have a dedicated ELT instance or reporting or sandbox instances, if required. These are all managed through EsgynDB Manager. This separation also means that you can scale storage and compute independently and that you can support multi-tenancy within a single instance of the database.

Why should you care?

There are a lot of database vendors addressing the market for hybrid transactional and analytic environments, most of whose solutions are proprietary. Relatively few leverage Hadoop and open source technologies to the extent that EsgynDB does. This has the obvious advantages, not only of cost, but also the ability to deploy complementary technologies such as HCatalog, Zookeeper, Yarn and so forth. The product’s high concurrency capabilities are a major differentiator, especially within the context of Hadoop-based implementations in general.

The Bottom Line

Esgyn, as a company, will be relatively new to most people. However, it should be treated with greater respect. Given its background and history, it is no surprise that EsgynDB is robust and richly featured, and it is strong contender for supporting hybrid processing environments, especially where a Hadoop-based implementation is preferred.

Commentary

Open Source Software Integrity

Solutions

EsgynDB

Research

Hybrid real-time data processing

There are multiple ways in which databases that support real-time operational/transactional and analytic processing may be implemented. This report compares representative samples of each.

EsgynDB

EsgynDB is a massively parallel processing SQL-based database that typically leverages Trafodion tables, but which also supports Hive, HBase, Parquet, ORC and Avro formats.

Further Information »