## 3. RDMA Architecture

### 3.1 RDMA Capabilities
```java
public enum RdmaCapability {
    RDMA_READ,           // Remote direct read operations
    RDMA_WRITE,          // Remote direct write operations  
    RDMA_SEND_RECEIVE,   // Traditional send/receive with RDMA
    MEMORY_REGISTRATION, // Dynamic memory registration
    FAST_REGISTRATION   // Fast memory region registration
}

participant dep_func as Dep scope="function"
participant operation as Path Operation

    client ->> dep_req: Start request
    Note over dep_req: Run code up to yield
    dep_req ->> dep_func: Pass dependency
    Note over dep_func: Run code up to yield
    dep_func ->> operation: Run path operation with dependency
    operation ->> dep_func: Return from path operation
    Note over dep_func: Run code after yield

        /**
         * Get the source node for move operations
         *
         * @return source node name
         */
        public String getSourceNode() {
            return sourceNode;
        }

        /**
         * Sets the source node for move operations.
         *
         * @param sourceNode the name of the source node in a share move operation
         */

 * This interface provides common functionality for requests that need to specify
 * and validate remote repositories for artifact resolution, dependency collection,
 * model building, and other Maven operations.
 *
 * <p>Implementations of this interface can specify a list of remote repositories
 * to be used during the operation. If no repositories are specified (null),

   * in a single load operation.
   */
  public double missRate() {
    long requestCount = requestCount();
    return (requestCount == 0) ? 0.0 : (double) missCount / requestCount;
  }

  /**
   * Returns the total number of times that {@link Cache} lookup methods attempted to load new
   * values. This includes both successful load operations and those that threw exceptions. This is

   * in a single load operation.
   */
  public double missRate() {
    long requestCount = requestCount();
    return (requestCount == 0) ? 0.0 : (double) missCount / requestCount;
  }

  /**
   * Returns the total number of times that {@link Cache} lookup methods attempted to load new
   * values. This includes both successful load operations and those that threw exceptions. This is

* the contents your program gave to the system to be written to disk
* a remote API operation
* a database operation to finish
* a database query to return the results
* etc.

As the execution time is consumed mostly by waiting for <abbr title="Input and Output">I/O</abbr> operations, they call them "I/O bound" operations.

   * changes the contents of the transposed table and vice versa.
   *
   * <p>The returned table supports update operations as long as the input table supports the
   * analogous operation with swapped rows and columns. For example, in a {@link HashBasedTable}
   * instance, {@code rowKeySet().iterator()} supports {@code remove()} but {@code

│   ├── main.py
│   └── test_main.py
```

Let's say that now the file `main.py` with your **FastAPI** app has some other **path operations**.

It has a `GET` operation that could return an error.

It has a `POST` operation that could return several errors.

Both *path operations* require an `X-Token` header.

{* ../../docs_src/app_testing/app_b_an_py310/main.py *}

        // of executing. Don't null out the latestTaskQueue field, because the work done here
        // may have scheduled more operations on another thread, and if those operations then
        // trigger reentrant calls that thread will have updated the latestTaskQueue field, and
        // we'd be interfering with their operation.
        executingTaskQueue.thread = null;
      }
    }

    private boolean trySetStarted() {