feat: Introduced a kafka-like `ShardingStrategy` for `inmemory`

feat: Introduced a kafka-like `ShardingStrategy` for `inmemory`

- Introduced the `ShardingStrategy`, that picks a shard for a given
  `ChatRoom`-ID.
  - Implemented a `KafkaLikeShardingStrategy`, that reuses the hashing
    algorithm, that is implementd in `Utils.murmur2()` in the
    `org.apache.kafka:kafka-clients` library.
  - The attribute `shard` of the `ChatHome` has to be restored according
    to the configured `ShardingStrategy` when loading the state - it must
    not be safed with the stored data, because it might change due to
    configuration-changes.
- The `ChatBackendController` was not configured correctly, because it
  had consumed the single `ChatHome` from the old configuration as the
  only entry in its `ChatHome[]`-array.
  - Refined the application-properties: Introduced an inner subclass
    `InMemoryServicesProperties` of `ChatBackendProperties`, that
    encapsulates the properties, that only concern the implementation
    `inmemory`.
  - Added the configuration-parameters `numShards` and `ownedShards`,
    that are needed by `inmemory`, to handle the sharding correctly.
- Introduced `ChatHomeFactory`, because the `ChatHome`s are instanciated
  by `ChatBackendConfiguration`, which is not aware of the configured
  implementation.
- Adjusted the test-cases to the changes.
- Added `InMemoryWithFilesAndShardingConfigurationIT`, that asserts,
  that the application works as expected, if sharding is activated.