InterSystems

Fig 2 - Smart data functional components

The IRIS database is a hybrid multi-model database. This allows it to manage a rich variety of different models and data types such as time series, documents etc. There is a federated query engine to allow queries to be distributed and executed at source systems where appropriate, very much in the data fabric philosophy.

Of course, in the real world, some data will end up being reused, and so there may be a need for caching of data for efficiency reasons. IRIS can also store data directly, so a customer can choose what data is accessed within IRIS and what data is accessed remotely, a choice that will be driven partly by query execution efficiency. The product has a reputation for strong scalability, a recurring theme in dozens of customer testimonials that I examined. The US Department of Veterans Affairs uses IRIS across 3 petabytes of data running on AWS, for example. One US healthcare provider uses IRIS to combine 300 million data elements from 56,000 healthcare providers, generating 9 million alerts a month related to patients or medical tests.

Within the IRIS catalogue, there is a capability to build cloud-based data services, which can be exposed to other applications. These data services can be combined into specific applications, and indeed IRIS has extensive application-building capability and several partners that build applications using it for various vertical applications in healthcare and finance, and also in the supply chain field. There is no explicit knowledge graph capability within the product, though given the nature of the IRIS database, third party partners have built such a tool on top of the core database.  IRIS includes visual tools designed for data stewards to easily manage data and semantic models. There is also a data pipeline management tool that has a dashboard for scheduling.

InterSystems has a natural language interface and has built AI chatbots that are specific to certain industries. They have a data transformation language of their own, and they are building an AI co-pilot product for this.

IRIS has a broad range of capabilities that enable its use as the basis of a modern data fabric architecture. It has a distributed query capability, a data catalogue, a flexible database architecture, an integration platform, a semantic layer and AI-driven analytics, all embedded in the InterSystems IRIS data platform. There is probably no single vendor that provides every single element of a complete data fabric architecture, but InterSystems provide many of the most important elements.

The bottom line

InterSystems has a very well-proven product that provides many key components of a modern data fabric architecture. Its strengths in scalability and rich set of partner applications make it a serious contender if you are considering implementing a data fabric architecture.

InterSystems IRIS stores data in multi-dimensional arrays. It supports relational, object (with full persistence, polymorphism, inheritance, and so forth, and with no requirement for object-relational mappings), document, and multi-dimensional models, and you can implement any number of these within the same environment, with full interoperability across these projections, and without any duplication of data. Note that you can read an array either vertically or horizontally, which means that you only need to store data once to support both transactions and analytics. This is a major differentiator for InterSystems.

InterSystems IRIS scales to accommodate large workloads and data sets on commodity hardware, with both database and application servers scaling out horizontally. Application code is decoupled from the persistence of data, which allows application servers to scale horizontally, independently of the number of shard servers, and to distribute workloads automatically, thereby supporting both performance and consistency. The environment is configured so that nodes are designated as query, transactional or hybrid servers, in order to cater for different workload resource requirements. Data ingestion can be parallelised directly to each shard server, providing high-performance ingestion for streaming data, and analytic queries can be pushed down to partitioned or sharded data tables, further increasing performance and resource efficiency. The software can also make direct use of graphics processors for pipelining and so forth.

Analytics capabilities provided by InterSystems IRIS include a new IntegratedML feature that allows you to create and use predictive models using automated SQL functions; support for PMML (Predictive Modelling Mark-up Language) and a server side Python runtime engine that allows native execution of predictive models; and a connector for leveraging Apache Spark-based machine learning and predictive models within the InterSystems IRIS environment (with parallel operations and high-speed connections from each of the shard servers into a Spark cluster). ‘Adaptive Analytics’ is another new feature that allows you to expose analytic data only once, in such a way that it can serve multiple use cases simultaneously. This is accomplished using a ‘virtual cube’ data model, an alternative to the OLAP cube, that can be assembled using a drag and drop interface and deployed to various business intelligence and visualisation tools, such as Tableau, Power BI, and Qlik (supported via an ODBC interface).

On that note, to facilitate the embedding of real-time business intelligence into operational applications, InterSystems IRIS includes a designer for creating dashboards; an analysis component, that can be employed by business users to explore and display relevant data; and an architect component, used to define your data model. For unstructured data, InterSystems IRIS includes natural language capabilities, and it supports Apache UIMA (Unified Information Management Architecture).

InterSystems IRIS provides several language options for development purposes. Apps that run on the platform directly can leverage SQL, Python, or ObjectScript (an in-house programming language) while external applications can also use Java, .Net or Node.js. Access to Python is a recent development and is a significant step forward in terms of access to trained developers (Python is more commonly used than ObjectScript) and programming libraries, without needing to sacrifice the performance, security, scalability and other benefits of an embedded approach. Notably, Python can run in the kernel directly on the data, and is considered a ‘full peer’ to ObjectScript, meaning that the two can essentially be treated as interchangeable and cross-compatible within InterSystems IRIS.

Hybrid real-time use cases are common in financial services (for instance, fraud detection). With a dedicated Financial Services Solution Suite, InterSystems IRIS caters to these environments. We consider InterSystems IRIS to be highly performant and scalable. It is, for instance, quite capable of processing transactions, indexing incoming data, and performing analytics on both real-time data and non-real-time data (that is, historical data and reference data) at scale and in real-time.

The platform also takes pains to make sure its analytics are not just highly effective in theory, but easy to build, access and use in practice. You can see this in practice with the specifically focused Business 360, Customer 360, InterSystems TotalView for Asset Management, and Cloud Fintech Gateway use cases.

The Bottom Line

Recent developments to InterSystems IRIS, the development of a smart data fabric and introduction of InterSystems Financial Services Solution Suite make for a compelling hybrid transactional/analytic offering for all financial institutions.

InterSystems IRIS stores data in multi-dimensional arrays. It supports relational, object (with full persistence, polymorphism, inheritance and so forth, and with no requirement for object-relational mappings), document, and multi-dimensional models, and you can implement any number of these within the same environment, with full interoperability across these projections, and without any duplication of data. Note that you can read an array either vertically or horizontally, which means that you only need to store data once to support both transactions and analytics. This is a major differentiator for InterSystems.

InterSystems IRIS scales to accommodate large workloads and data sets on commodity hardware, with both database and application servers scaling out horizontally. Application code is decoupled from the persistence of data, which allows application servers to scale horizontally, independently of the number of shard servers, and to distribute workloads automatically, thereby supporting both performance and consistency. The environment is configured so that nodes are designated as query, transactional or hybrid servers, in order to cater for different workload resource requirements. Data ingestion can be parallelised directly to each shard server, providing high-performance ingestion for streaming data, and analytic queries can be pushed down to partitioned or sharded data tables, further increasing performance and resource efficiency. The software can also make direct use of graphics processors for pipelining and so forth.

Fig 1 - InterSystems IRIS adaptive analytics

Analytics capabilities provided by InterSystems IRIS include a new IntegratedML feature that allows you to create and use predictive models using automated SQL functions; support for PMML (Predictive Modelling Mark-up Language) that allows native execution of predictive models developed in R, IBM SPSS and other environments; and a connector for leveraging Apache Spark-based machine learning and predictive models within the InterSystems IRIS environment (with parallel operations and high-speed connections from each of the shard servers into a Spark cluster). ‘Adaptive Analytics’ is another new feature – see Figure 1 – that allows you to expose analytic data only once, in such a way that it can serve multiple use cases simultaneously. This is accomplished using a ‘virtual cube’ data model, an alternative to the OLAP cube, that can be assembled using a drag and drop interface and deployed to various business intelligence and visualisation tools, such as Tableau, Power BI and Qlik (supported via an ODBC interface).

On that note, to facilitate the embedding of real-time business intelligence into operational applications, InterSystems IRIS includes a designer for creating dashboards; an analysis component, that can be employed by business users to explore and display relevant data; and an architect component, used to define your data model. For unstructured data, InterSystems IRIS includes natural language capabilities, and it supports Apache UIMA (Unified Information Management Architecture).

InterSystems IRIS provides several language options for development purposes. Apps that run on the platform directly can leverage SQL, Python, or ObjectScript (an in-house programming language) while external applications can also use Java, .Net or Node.js. Access to Python is a recent development and is a significant step forward in terms of access to trained developers (Python is more commonly used than ObjectScript) and programming libraries, without needing to sacrifice the performance, security, scalability and other benefits of an embedded approach. Notably, Python is considered a ‘full peer’ to ObjectScript, meaning that the two can essentially be treated as interchangeable and cross-compatible within InterSystems IRIS.

Hybrid real-time use cases are common in financial services (for instance, fraud detection), retail (customer personalisation), manufacturing/IoT (detecting errors, predictive maintenance), and various other industries. InterSystems IRIS not only caters to these environments, but it is going to be more cost efficient than many of its rivals, thanks to the fact that you only need to store your data once and in only one place.

Moreover, we consider InterSystems IRIS to be highly performant and scalable. It is, for instance, quite capable of processing transactions, indexing incoming data, and performing analytics on both real-time data and non-real-time data (that is, historical data and reference data) at scale and in real-time.

The platform also takes pains to make sure its analytics are not just highly effective in theory, but easy to build, access and use in practice. This is described by InterSystems as the ‘last mile’ of analytics, and positioned as an essential capability for generating real value from analytics technology. We are inclined to agree.

The Bottom Line

We are long-time fans of InterSystems products, and the recent improvements to InterSystems IRIS have certainly not dampened our enthusiasm: it remains a compelling hybrid transactional/analytic offering.

InterSystems IRIS stores data in multi-dimensional arrays. It supports relational, object (with full persistence, polymorphism, inheritance, and so forth, and with no requirement for object-relational mappings), document, and multi-dimensional models, and you can implement any number of these within the same environment, with full interoperability across these projections, and without any duplication of data. Note that you can read an array either vertically or horizontally, which means that you only need to store data once to support both transactions and analytics. This is a major differentiator for InterSystems.

InterSystems IRIS scales to accommodate large workloads and data sets on commodity hardware, with both database and application servers scaling out horizontally. Application code is decoupled from the persistence of data, which allows application servers to scale horizontally, independently of the number of shard servers, and to distribute workloads automatically, thereby supporting both performance and consistency. The environment is configured so that nodes are designated as query, transactional or hybrid servers, in order to cater for different workload resource requirements. Data ingestion can be parallelised directly to each shard server, providing high-performance ingestion for streaming data, and analytic queries can be pushed down to partitioned or sharded data tables, further increasing performance and resource efficiency. The software can also make direct use of graphics processors for pipelining and so forth.

Analytics capabilities provided by InterSystems IRIS include a new IntegratedML feature that allows you to create and use predictive models using automated SQL functions; support for PMML (Predictive Modelling Mark-up Language) and a server side Python runtime engine that allows native execution of predictive models; and a connector for leveraging Apache Spark-based machine learning and predictive models within the InterSystems IRIS environment (with parallel operations and high-speed connections from each of the shard servers into a Spark cluster). ‘Adaptive Analytics’ is another new feature that allows you to expose analytic data only once, in such a way that it can serve multiple use cases simultaneously. This is accomplished using a ‘virtual cube’ data model, an alternative to the OLAP cube, that can be assembled using a drag and drop interface and deployed to various business intelligence and visualisation tools, such as Tableau, Power BI, and Qlik (supported via an ODBC interface).

On that note, to facilitate the embedding of real-time business intelligence into operational applications, InterSystems IRIS includes a designer for creating dashboards; an analysis component, that can be employed by business users to explore and display relevant data; and an architect component, used to define your data model. For unstructured data, InterSystems IRIS includes natural language capabilities, and it supports Apache UIMA (Unified Information Management Architecture).

InterSystems IRIS provides several language options for development purposes. Apps that run on the platform directly can leverage SQL, Python, or ObjectScript (an in-house programming language) while external applications can also use Java, .Net or Node.js. Access to Embedded Python is a recent development and is a significant step forward in terms of access to trained developers (Python is more commonly used than ObjectScript) and programming libraries, without needing to sacrifice the performance, security, scalability and other benefits of an embedded approach. Notably, Python can run in the kernel directly on the data and is considered a ‘full peer’ to ObjectScript, meaning that the two can essentially be treated as interchangeable and cross-compatible within InterSystems IRIS.

InterSystems Supply Chain Orchestrator offers, not only the ability to provide powerful analytics, on a public cloud or on-premises, using data from existing systems, but also as an underlying transactional platform for newly developed microservices-based applications to run on. InterSystems has a number of publicly available reference cases, like SPAR Austria, a member of SPAR, the world’s largest food retailer consortium, which is a €4 billion company with more than 800 outlets and 600 SPAR merchants in Austria. SPAR Austria initially developed a complete warehouse management system utilising the InterSystems data platform for SPAR stores across Eastern Europe. This has now been extended to 1400 stores in Austria and is planned to be rolled out into Italy as well.

While SPAR used the IRIS data platform to develop, very rapidly, new custom applications, UST, a leading digital transformation solutions company required a different approach. It has integrated its Optum solution, based on SAP/Hana, with Supply Chain Orchestrator to deliver an Azure cloud-based Optimisation-as-a-Service that achieves enhanced supply chain orchestration and gains end-to-end visibility.

Such a hybrid real-time platform offers the opportunity to develop quickly new use cases to meet the never-normal environment of the 2020s. We consider the underlying InterSystems IRIS platform to be highly performant and scalable. It is, for instance, quite capable of processing transactions, indexing incoming data, and performing analytics on both real-time data and non-real-time data (that is, historical data and reference data) at scale and in real-time.

The platform also takes pains to make sure its analytics are not just highly effective in theory, but easy to build, access and use in practice. You can see this in practice with the specifically supply chain focused canonical data model and analytics cubes.

The Bottom Line

Recent developments to InterSystems IRIS, the development of a smart data fabric and introduction of InterSystems Supply Chain Orchestrator solution make for a compelling hybrid transactional/analytic offering for the management and optimisation of supply chains.