If you missed the amazing, first Apache Geode Summit held this past March, never fear, all of the slides and video replays of the sessions have been posted for your viewing pleasure.
We can describe this content in just a few words—amazing talks, excellent speakers, deep experience.
And, here are three big things you can learn by checking out the replays:
- #1: Apache Geode Scales Mission Critical Financial Services & Travel Apps
- #2: Geode Powers Modern, Data-Driven Architectures
- #3: Geode Has a Promising Future Ahead
Each of these are 30 minute sessions. So, it is easy to make some time and check them out. Microwaving a bowl of popcorn before viewing is recommended to maximize your enjoyment.
Takeaway #1: Geode Scales Mission-Critical Financial Services & Travel Apps
We heard from a number of production users about their high scale, high performance applications which take advantage of Geode’s distributed in-memory processing.
Combining Stream Processing and In-Memory Data Grids for Near-Real Time Aggregation and Notifications
First, Olivier Mallassi presented how Murex uses Geode in their risk management solution for financial services. In his case, Apache Storm and Geode are integrated, even collocated, and applied to the fast processing side of a lambda architecture. While Geode is used for immutable events and event logging, one of the most powerful features he covers is continuous queries. Developers can think of these as a distributed, scalable set of in-memory listeners which use data to identify and push event notifications to clients.
Slides can be found here.
Wall St. Derivative Risk Solutions Using Geode
Then we heard from Andre Langevin, currently with Canadian Imperial Bank of Commerce, who has built several such systems using the technology in Apache Geode. In his talk, he shares a series of insightful architecture whiteboards, explaining how Geode is used as a data AND compute grid, rolling up financial risk across multiple trading systems.
Slides can be found here.
Design Tradeoffs in Distributed Systems—How Southwest Airlines Uses Geode
Finally, we heard from Brian Dunlap of Southwest Airlines. He explained just how Geode provides real-time intelligence, integrating data from reservation feeds, checkins, kiosks, boarding, gates, flights—basically from everything that moves. After explaining how his system supports critical operations and decisions, regularly crunching out optimizations for over 1,000,000 schedules, he shares many of the lessons learned.
Slides can be found here.
Takeaway #2: Geode Powers Modern, Data-Driven Architectures
If the previously mentioned talks didn’t give you a sense of Geode’s fit in data-driven architectures, we also heard from a number of architects on design patterns for Geode in modern applications.
Where Does Geode Fit in Modern System Architectures?
With years of experience using Geode, Eitan Suez explained how developers and architects, like him, get started with Geode and realize that it can be used to solve a lot of problems—that it is basically a distributed hash map. He then presented an approach to using Geode as part of a CQRS (Command Query Responsibility Segregation) architecture.
Slides can be found here.
Architecting Data-Driven, Smarter Cloud Native Apps with Real-Time Decision-Making
Next, we heard from Fred Melo. He presented a compelling explanation and demo of HTAP (Hybrid Transactional / Analytical Processing) architectures, with real-time predictive analytics, featuring Apache Geode.
Slides can be found here.
Large Scale Fraud Detection Using GemFire Integrated with Greenplum
Finally, we heard about a fraud detection case study where Apache Geode was deployed as a pre-processing engine to detect events in streaming data before being loaded into a data warehouse. This included a business-logic processing layer for custom applications above the database. This particular application used the GemFire Connector to Greenplum Database.
Slides can be found here.
Takeaway #3: Geode Has a Promising Future Ahead
In another track of the Summit, many of the committers and contributors of Apache Geode convened to discuss the current and future state of Apache Geode’s feature set. When you look at the list of capabilities below, you can see that Apache Geode as some very compelling potential features in its future.
- Spring Data GemFire API current & future
- Integration & Future Direction for Spring Cloud Data Flow & Geode
- Redis to Geode Adaptor
- Off-Heap Storage Current and Future Design
- Integration with Spark – Geode reading and writing RDDs
Many of these features are still in their conceptual phase, and the Apache Geode community is seeking people interested to help contribute to make these features better. We finished up with a great talk by Nitin Lamba who described his journey to becoming an Apache Geode committer: Easy Ways to Become a Contributor to Apache Geode. You can also join the mailing list.
About the Author
