Added diagram for the new lake tiering service architecture

Author: wuchong

website/blog/assets/0.7/new_tiering_service.png

@@ -63,6 +63,9 @@ The sync offset is now persisted directly within Fluss metadata, with future pla
 Furthermore, Flink jobs have transitioned from maintaining persistent table subscriptions to a more efficient model where they process one table at a time,
 dynamically requesting the next table upon completion of the current one.
 
+
+![New Architecture Diagram](../assets/0.7/new_tiering_service.png)
+
 This optimization significantly reduces the load on Fluss and boosts batch processing efficiency. As a result, cluster operators gain:
 * **Enhanced Service Robustness:** Operators can launch multiple Flink jobs to distribute workload and increase overall system resilience.
 * **Flexible Resource Management:** The ability to stop idle jobs at any time allows for immediate reclamation of valuable cluster resources.