Sign in to confirm you’re not a bot
This helps protect our community. Learn more

How we used Azure Cosmos DB to power AEP's Datagroups

Microsoft Developer
588K subscribers
<__slot-el>
<__slot-el>
64 views 2 years ago #AzureCosmosDB
This talk will cover the key features of Azure Cosmos DB which helped us build Datagroup feature. We were required to build an object hierarchy container to support Adobe Experience Platform's Datagroups used for Dashboard and Reporting. While we wanted a document stores to support search and sorting requirements, we also had to build a hierarchical model (similar to files and folder). That presented it's own unique challenges. Given that a single object may belong to multiple datagroups, deletion of a single object would have required updating several containers which cannot be done atomically on most Document stores. Also, a Graph Database presented scalability issues. As a result both Document stores and Graph databases were unsuitable for this usecase, and we were looking for something in middle, and that's where Azure Cosmos DB's unique capability of combining Document Database with Graph API helped us build this feature which scaled horiz …...more
...more
How we used Azure Cosmos DB to power AEP's Datagroups
0Likes
64Views
2023Mar 28
This talk will cover the key features of Azure Cosmos DB which helped us build Datagroup feature. We were required to build an object hierarchy container to support Adobe Experience Platform's Datagroups used for Dashboard and Reporting. While we wanted a document stores to support search and sorting requirements, we also had to build a hierarchical model (similar to files and folder). That presented it's own unique challenges. Given that a single object may belong to multiple datagroups, deletion of a single object would have required updating several containers which cannot be done atomically on most Document stores. Also, a Graph Database presented scalability issues. As a result both Document stores and Graph databases were unsuitable for this usecase, and we were looking for something in middle, and that's where Azure Cosmos DB's unique capability of combining Document Database with Graph API helped us build this feature which scaled horizontally to 1000s of users We leveraged Azure Cosmos DB's unique features such Stored Procedure to ensure atomicity in operation when a container gets updated with multiple children object. By using Gremlin queries, we were able to compute graph complexity and enforce restrictions to throttle any usage, to avoid long traversals. The most interesting part of using Azure Cosmos DB is that we are charged for throughput and not for capacity. That allowed us to prototype various models and then scale the traffic ramps up and thus avoiding the need for provisioning additional capacity that may never be used. https://aka.ms/AzureCosmosDBConf/Parag #AzureCosmosDB

Follow along using the transcript.

Microsoft Developer

588K subscribers