The following section explains how a diet_call is managed when used on the previous architecture.
If a client sends a diet_call for the problem a to the alpha's MA, the alpha's MA will return a reference of one of it's SeD. However, if its scheduler (see Section 8) says that no SeD is available, it will forward the request to beta and gamma. If beta has an available SeD, it will be used to resolve the problem. If not, the request is forwarded to delta.
Now, if a client performs a diet_call for the problem c to the delta's MA, the delta MA does not have a SeD that can resolve this problem. So, it forwards the request to gamma. If gamma has no available SeD, the request is forwarded to alpha and beta.