# Apache Geode

{% hint style="info" %}
Fractalworks has extensive experience developing real-time event based solutions and products that applies this technology. 

Please [contact us](https://www.fractalworks.io/contact) for for details on how we can help you.
{% endhint %}

## Overview

Apache Geode integrates with Joule as a **high-performance, in-memory key-value store**. As an essential part of the the platform, it provides **low-latency access** to contextual data for advanced stream processing.

Known for its **reliability and scalability**, Geode has supported large-scale enterprise solutions for over 20 years.

In Joule, Geode enables **immediate updates** from a distributed cluster, allowing processors to use the latest data in real-time without manual refreshes. To learn more about the features of Apache Geode, we invite you to read this [article](https://boglesby-2508.medium.com/what-is-stored-in-an-apache-geode-region-55ee0093e574) for a really good overview. 

{% hint style="info" %}
**Driver details:** [org.apache.geode:geode-core:1.15.1](https://geode.apache.org/)
{% endhint %}

Storage regions can be **primed on startup** using either a custom query or a `getAll` function otherwise data is fetched on cache misses.

Furthermore, for cached data elements, data updates are immediately propagated from the connected distributed cluster to the Joule process thereby enabling up-to-date contextual data.

## Use cases

Geode is ideal for handling various types of static or slow-moving reference data, critical in scenarios requiring low-latency access and high concurrency. Common examples include:

Typical use cases include:

1. <mark style="color:green;">**Geographic and regional data**</mark>\
   Postal codes, country codes (e.g., ISO-366).
2. <mark style="color:green;">**Device and market data**</mark>\
   Mobile manufacture models, market exchange codes.
3. <mark style="color:green;">**Currency data**</mark>\
   Currency codes for international transactions.
4. <mark style="color:green;">**Machine learning models**</mark>\
   Pre-trained models for real-time predictions.
5. <mark style="color:green;">**Static variables and codes**</mark>\
   Charge codes, car VINs, pre-computed data for analysis.
6. <mark style="color:green;">**Pattern matching**</mark>\
   Regex patterns for validating inputs, such as phone numbers.

### Architecture

Joule leverages a connected Geode client cache that initially links through the Geode locator, and then after connects directly to cluster members. This approach **supports on-demand loading** of dynamic contextual data, enhancing data retrieval speeds and ensuring scalability for high-throughput environments.

<figure><img src="https://3062398388-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FUU6FZlV07ZD90OzbzGww%2Fuploads%2FwbnoMRnvOg1JxwLm3HA5%2Fgeode_arch.png?alt=media&#x26;token=7af7448b-3b63-48f6-bbcb-8aaed503c661" alt=""><figcaption><p>Joule integration architecture for Apache Geode</p></figcaption></figure>

It’s recommended to review [Apache Geode documentation](https://geode.apache.org/) to understand the distributed data cluster architecture. This will inform you of the benefits of the solution.

## Example & DSL attributes

The following example is a configuration for integrating Apache Geode as a data source within Joule:

```yaml
contextual data:
 ...
    - geode stores:
          name: us markets
          connection:
              locator address: 192.168.86.39
              locator port: 41111
          stores:
            nasdaqIndex:
              region: nasdaq-companies
              keyClass : java.lang.String
              gii: true
            holidays:
              region: us-holidays
              keyClass : java.lang.Integer
```

### **Explanation**

This configuration specifies **Geode** as a data source, linking two key regions: `nasdaq-companies` and `us-holidays`.

The `gii` (get-initial-image) option for `nasdaqIndex` ensures data is primed on startup, minimising cache misses. By connecting via the provided locator address and port, Joule gains access to the Geode distributed cluster, ensuring the most up-to-date data in the processing pipeline.

### Attributes schema

<table><thead><tr><th width="171">Attribute</th><th width="280">Description</th><th width="188">Data Type</th><th data-type="checkbox">Required</th></tr></thead><tbody><tr><td>name</td><td>Logical name of reference data set for given target connection</td><td>String</td><td>true</td></tr><tr><td>locator address </td><td>Locator process IP address</td><td>String<br>Default: localhost</td><td>true</td></tr><tr><td>locator port</td><td>The port the cluster locator listens for cluster and client connections</td><td>Integer<br>Default: 41111</td><td>true</td></tr><tr><td>stores</td><td>Map of logical names to region configurations. See Region attributes</td><td>List of logical storage name to Geode storage region</td><td>true</td></tr></tbody></table>