summaryrefslogtreecommitdiff
path: root/docs/architecture.md
diff options
context:
space:
mode:
Diffstat (limited to 'docs/architecture.md')
-rw-r--r--docs/architecture.md25
1 files changed, 25 insertions, 0 deletions
diff --git a/docs/architecture.md b/docs/architecture.md
new file mode 100644
index 00000000..032e8f4b
--- /dev/null
+++ b/docs/architecture.md
@@ -0,0 +1,25 @@
+# Architecture
+
+OpenDC consists of four components: a Kotlin simulator, a MongoDB database, a Python
+Flask [API](/opendc-web/opendc-web-api), and a React.js [frontend](/opendc-web/opendc-web-ui).
+
+<p align="center">
+ <img src="https://raw.githubusercontent.com/atlarge-research/opendc/master/misc/artwork/opendc-component-diagram.png" alt="OpenDC Component Diagram">
+</p>
+
+On the frontend, users can construct a topology by specifying a datacenter's rooms, racks and machines, and create
+scenarios to see how a workload trace runs on that topology. The frontend communicates with the web server over
+SocketIO, through a custom REST request/response layer. For example, the frontend might make a `GET` request
+to `/api/v1/users/{userId}`, but this request is completed via SocketIO, not plain HTTP requests. Note that the API
+itself can also be accessed by HTTP.
+
+The (Swagger/OpenAPI compliant) API spec specifies what requests the frontend can make to the web server. To view this
+specification, go to the [Swagger Editor](https://editor.swagger.io/) and paste in
+our [opendc-api-spec.yml](../opendc-api-spec.yml).
+
+The web server receives API requests and processes them in the database. When the frontend requests to run a new
+scenario, the web server adds it to the `scenarios` collection in the database and sets its `state` as `QUEUED`.
+
+The simulator monitors the database for `QUEUED` scenarios, and simulates them as they are submitted. The results of the
+simulations are processed and aggregated in memory. Afterwards, the aggregated summary is written to the database, which
+the frontend can then again retrieve via the web server.