guest@ldsoftware.it:
  1. ~
  2. /software
  3. /ragno
  4. /architecture
$
Back to the home page

Architecture

The architecture of Ragno is based on microservices and eventsourcing. In this page I will briefly present the blueprint of the whole system, explaining what role each part plays in the whole.

Let’s start by showing the basic diagram of the components that make up the whole.

(TODO put an image)

The event storage

The event storage is the most important part of the system, as it will be the single source of truth of the system, and it will contain all the data relative to the sites. It may sound counter-intuitive to have a distributed environment where every service relies on a single data storage, but in the event-sourcing pattern this is the norm: the event storage will be the main faucet from where all the service will obtain data, and then each one of them will use that data to create its own private read model that will allow it to be independent from the other services and (if needed) from the single source of truth itself.

The event storage has the double function of being a data storage and a message bus, since some services may be interested in some events only for triggering background jobs.

The choice for the event storage, after much debate, went to Event Store. Initially - a long long time ago when my journey into microservices had just begun - I was suggested using Kafka, but

  • I’ve never found it attractive, because of its high complexity
  • It is NOT the correct choice for an event store (in fact, it is a messaging system)

The site manager

Moving down from the “city center”, we encounter the first component, the most important one at that, that is the site manager.

The site manager is the entry-point of the websites domain, and will emit all the events that pertain to the sites domain (see the domains page for more information).

As anticipated, it won’t be your classical CRUD application, as it will not generate data that goes in a table, but events that will go directly to the event store. It will nevertheless contain a read-model that will help taking decisions, e.g. if the current user has the permission to act on a certain domain, or stuff like that.

The site builder

The site builder is another component, but is not exposed to the network. It will take care of building the websites based on the events that run into the main event storage. The output of the build will be based on the website’s configuration, so this component will rely on read models to make up decision tables; and the website itself is a read model too.

The indexer

The site indexer is the third component that I plan to introduce in the system, and it’s another read-model that will listen to the events running through the system and create a full-text search engine for the website pages. It will expose a classical REST API that will allow the live site to query for its contents (e.g. via a search bar) or any third party that wants to query the site.

The UI

Finally, the UI is the only interaction point between the user and Ragno. It will allow easy editing and managing of the websites.