API Modules

Build Application-Defined Data Structures with API Modules

Today’s data structure stores boost cache performance by accessing cached objects as data structures, like hashed or sorted sets, instead of as uninterpreted data (BLOBs). However, it’s a huge challenge to create and deploy new types of data structures in the cache.

Enter API modules, application-defined code that you can develop in Java and C# and then dynamically deploy to ScaleOut StateServer. With API modules, you create a strongly typed data structure for accessing a namespace of cached objects and then define APIs that clients can call to manage these objects. API modules let you quickly and easily implement application functions and reduce data motion to and from the cache.

Here’s an example of a shopping cart data structure that could be created by an e-commerce company to provide custom functionality for its shopping site:

Because API modules can be developed and deployed as the need arises, it’s practical to create specialized APIs for each application. For example, the e-commerce company might create APIs that total cart items by brand or product type. These APIs run efficiently within the cache and minimize data motion. They also take advantage of parallel speedup since API modules are deployed to all cache servers.

Message Modules

Streamline Event-Processing with Message Modules

Deploy Java or C# application code called message modules in the distributed cache to receive and handle incoming messages. These modules provide reliable, strongly typed access to cached objects. In-memory data storage accelerates performance by avoiding the overhead of accessing backing stores to retrieve needed data.

For example, consider an airline that stores flight and passenger information in a cloud-based, NoSQL database and uses an event-driven architecture to process system updates, like weather rerouting and equipment changes. Instead of using serverless functions to handle events, the airline can send incoming messages to a ScaleOut StateServer distributed cache and process them with message modules:

To accelerate processing, these modules immediately access cached objects when events flow in and use application-specific code to make changes. Objects move to and from the backing store as needed with “hot” data automatically staged in the cache on demand. Modules run in parallel on all cache servers to further speed up event processing.

Management UI

Dynamically Deploy Modules and Visualize Live Data

Unlike other data structure stores, ScaleOut StateServer is designed to let you deploy modules on the fly. Using the ScaleOut Active Caching Management UI running on a networked workstation, you can upload message and API modules to an operational distributed cache. The management UI lets you connect the cache to messaging hubs, like Kafka and Azure IoT Hub, as well as to a built-in REST service to receive messages. You can also connect modules to persistence stores, like SQL Server, AWS Dynamo DB, and Azure Cosmos DB:

The management UI provides facilities for aggregating and visualizing cached data that has been serialized using JSON. You can create charts which combine data and display it in groups, for example, total sales by region. You can also create both tabular and geospatial queries that display selected properties for cached objects.

Here are some examples:

Both charts and queries continuously update every few seconds to show the latest data. They use integrated map/reduce code to quickly compute results in parallel across all cache servers.

For more info, check out the ScaleOut Active Caching™ User Guide.