Software AG Apama

Apama is the crux of Software AG’s streaming analytics platform, and was the first product of its type to be introduced to the market. However, it is not the only component of Software AG’s streaming analytics solution. Under the covers, its analytics platform comprises several additional components including, but not limited to, Cumulocity Core for managing, integrating and connecting to IoT sensors and actuators; Zementis for deploying predictive models; the Industry Analytics Kit, which provides pre-built capabilities of various types; and MashZone NextGen, which is a visualisation and dashboarding product.

Apama itself is currently in version 10.1 and is available as both a proprietary Enterprise Edition and a freely available Community Edition. It can be deployed in-cloud, on-premises, or on IoT gateways and devices. It can also be deployed in a Docker or Cloud Foundry container.

Customer Quotes

“Equipment servicing costs are down an astounding 40 percent, and revenue is forecast to grow by half (thanks to implementing Apama, Terracotta DB and WebMethods)”
Global leader in image technology and print services

“The ability to respond quickly to client requests and roll out completely new service offerings in two months gave us a huge strategic advantage. Our team, working with Software AG’s IoT platform (Apama and WebMethods), made it happen.”
Royal Dirkzwager

Software AG Apama essentially has three elements: an integration framework for connecting to external systems and data feeds, an execution engine known as the Apama Correlator, and front-end tools for developers. As far as integration is concerned, Apama ships with connectivity to JMS, Universal Messaging, MQTT, Kafka, HTTP, databases, web services, files on disk and 27 capital markets-specific sources. There is also a connectivity plug-in framework that allows you to create your own in-process adapters and to reuse elements of these plug-ins across other adapters. For example, you can reuse JSON mapping logic across multiple adapters, so you only have to define such processes once.

Apama Correlator consists of multiple parallelised containers that are known as “contexts”. Multiple (thousands) of contexts may be active simultaneously and each context can process any number of Apama applications, which are written in either EPL (Apama’s event processing language) or Java (Apama has a built-in JVM). An optimised thread scheduler allows elastic scaling. Within the Correlator there are three major components: the Apama HyperTree, the Temporal Sequencer and the Stream Processor. Respectively, these provide an indexing method that allows low latency matching between events and patterns you are trying to detect; support for temporal and spatially-based correlations; and the ability to store and organise events that occur within a time interval and orchestrate real-time analytics across those time windows. Additionally, events can be stored in Software AG’s TerracottaDB, an in-memory NoSQL database. This is useful, for example, for long-running queries. Apama Correlator also features tight integration with Zementis for deploying predictive models in Apama, with support for both PMML (Predictive Modelling Mark-up Language) and TensorFlow models.

There are three front-end tools provided for developers of Apama applications, intended for use by developers, business analysts and end users respectively. The first of these is an Eclipse-based environment that supports development using EPL, the second provides a drag-and-drop interface wherein you graphically model declarative patterns across windows of data, and in the third case there is a web-based, wizard-driven interface whereby you parameterise pre-built templates. Software AG also offers a range of pre-built applications (accelerators) that span a variety of use cases, including IoT, intelligent GRC (governance, risk and compliance), and smart logistics.

Figure 1 – A dashboard created in MashZone NextGen

A variety of pre-built components and capabilities are available inside Apama via the Industry Analytics Kit. There are a number of commonly occurring quality issues that can arise with streaming events, as well as various fundamental analytics that are regularly reused, and what Software AG has done, is to pre-build these components for you. It has also developed extended capabilities that will, for example, link different capabilities together, such as linking a threshold breach to a missing data analytic that will look for any gaps in regular breaches. Software AG has also developed a number of proprietary, industry specific capabilities that are available in the Industry Analytics Kit.

Apama also integrates bi-directionally with MashZone NextGen, Software AG’s self-service analytics environment for preparing, visualising and analysing data. In particular, both Apama and MashZone NextGen can be used to build real time, interactive dashboards (an example of which can be seen in Figure 1).

Software AG Apama was the first product of its type to be introduced and it has been, and remains, a market leader. Software AG’s streaming analytics portfolio, Apama included, is extremely comprehensive and performant. In demonstration of the latter, Apama has been benchmarked (by Intel) with a throughput of approximately 70 million records per second on a single server but is capable of running on as little as an ARM v7 processor or a Raspberry Pi.

Moreover, Software AG is ahead of the curve when it comes to IoT. Apama is deployable on IoT gateways and at the edge, while Software AG’s Cumulocity IoT platform (which integrates with Apama) supports more than 350 protocols used by a variety of vendors of sensors, industrial devices and gateways. Where other vendors in the space are only beginning to see IoT deployments, Software AG already has over 300.

The Bottom Line

From a streaming analytics point of view, it is difficult to think of anything that you might require that is not provided within Apama and its surrounding products. Moreover, the Industry Analytics Kit and Software AG’s accelerators mean that you can get started with an Apama-based project much more quickly than you might expect. In summary, the company provides an impressive product set for streaming analytics.