Event sourcing is a model for handling software change. It allows the system to continuously create and complete events without going into chaos or inconsistency. Issues are typically resolved at their root cause instead of later, like an issue discovered but not fixed because the developer feared breaking their codebase (even though it was working fine).
Let’s consider an event sourcing architecture where the view layer consumes messages from multiple sources. Message inventory-bootstrapping logic must perform a database lookup for each message. Yet, each database lookup could be slow because it requires blocking for network access, disk writes, and locking to maintain consistency across clients.
In his seminal paper, “On the Use of Event Sourcing for Managing Application States,” Jayfaced wrote that database state management is inherently more scalable and fault tolerant than event sourcing. His argument is that events are ephemeral and cannot be used to reconstruct the past. On the other hand, a database can reproduce a past snapshot of its state at any given time.
Since then, many developers have adopted event sourcing as their state management technique. But is it the best option? In this article, we’ll explore why Jayfaced’s modeling assumption – that events are always reliable – may not hold in practice. We’ll look at ways you can compensate for potential faults in your system’s data, and show how to build a robust database-based state manager using MongoDB Atlas and Long-Term Revision History (LTRH).
What is Event Sourcing?
Event sourcing is a pattern that describes how an application can capture changes to the data it manipulates, typically through transactions. This allows for consistent querying and data modification throughout a system’s history.
A key benefit of event sourcing is that it can enable your application to respond reliably and consistently to changes. For example, if someone updates a record in an application using event sourcing, ensure that the change is reflected in all caching mechanisms and the database itself – users will not be able to see or interact with stale data if events are not properly captured and processed.
Event sourcing should only be used when appropriate due to its complexity. While event sourcing offers advantages over traditional methods, such as improved reliability, these advantages come at a cost. Event sourcing requires more work from developers to set up and manage and more storage space for historical information. Therefore, event sourcing may not be right for every application. In certain cases, such as small or web applications built on lightweight frameworks, it may be sufficient to rely on traditional methods instead of event sourcing.
Why use a Database?
A database is a valuable asset for storing and retrieving large amounts of data. There are lots of advantages to using a database for your event sourcing project, including the following:
- A database can scale to accommodate large amounts of data.
- A database can organize and store data in a manner that is easy to use.
- A database can help you create long-term storage solutions for your event-sourcing project.
Pros of an Event Sourcing architecture
Event sourcing is a great way to manage events, but it isn’t always the best option. A database can also be a good way to manage events. Here’s why: Event sourcing relies on a single source of truth. With a database, you have multiple sources of truth. If something goes wrong with the event store, you can still retrieve the events from the database and continue processing them as normal. Database-backed event sourcing also has other advantages:
- You can scale your event store more easily than with a single event source.
- You can use complex indexes and queries to find specific events.
- You can back up your data and restore it even if the event store crashes.
An Alternative Example: Oracle Events
There are a few alternatives to event sourcing if you don’t want to use Oracle. One option is to use a database. Another is to use a combination of databases and event sourcing.
A database can be used as a storage layer for events. In this scenario, the events would be stored in the database and the application would use the database to manage and query the events.
A combination of databases and event sourcing can store and manage events using an application. The application would store the events in a single data store and use the event streaming capabilities of the two separate databases to provide real-time updates of changes to the event data.
A database might be the right solution for you if you’re looking for an even more detailed way to manage your events. A database will give you more control and alignment over all aspects of your event life cycle, from planning and organizing to data management. Plus, it can make it easier to share data with other teams in your organization and react quickly to changing needs. If you’re considering switching to a database-based event life cycle management system, now is the time to do so!