diff --git a/ipc/MessageQueueMessageIF.h b/ipc/MessageQueueMessageIF.h
index 91753ced9..b5a30c086 100644
--- a/ipc/MessageQueueMessageIF.h
+++ b/ipc/MessageQueueMessageIF.h
@@ -1,24 +1,13 @@
 #ifndef FRAMEWORK_IPC_MESSAGEQUEUEMESSAGEIF_H_
 #define FRAMEWORK_IPC_MESSAGEQUEUEMESSAGEIF_H_
+
+#include <fsfw/ipc/messageQueueDefinitions.h>
 #include <cstddef>
 #include <cstdint>
 
-/*
- * TODO: Actually, the definition of this ID to be a uint32_t is not  ideal and
- * breaks layering. However, it is difficult to keep layering, as the ID is
- * stored in many places and sent around in MessageQueueMessage.
- * Ideally, one would use the (current) object_id_t only, however, doing a
- * lookup of queueIDs for every call does not sound ideal.
- * In a first step, I'll circumvent the issue by not touching it,
- * maybe in a second step. This also influences Interface design
- * (getCommandQueue) and some other issues..
- */
-
-typedef uint32_t MessageQueueId_t;
-
 class MessageQueueMessageIF {
 public:
-	static const MessageQueueId_t NO_QUEUE = -1;
+
     /**
      * @brief   This constants defines the size of the header,
      *          which is added to every message.
diff --git a/ipc/MessageQueueSenderIF.h b/ipc/MessageQueueSenderIF.h
index d9692f540..beb27f507 100644
--- a/ipc/MessageQueueSenderIF.h
+++ b/ipc/MessageQueueSenderIF.h
@@ -1,37 +1,26 @@
 #ifndef FRAMEWORK_IPC_MESSAGEQUEUESENDERIF_H_
 #define FRAMEWORK_IPC_MESSAGEQUEUESENDERIF_H_
 
+#include "../ipc/MessageQueueIF.h"
+#include "../ipc/MessageQueueMessageIF.h"
 #include "../objectmanager/ObjectManagerIF.h"
-class MessageQueueMessage;
-
-
-//TODO: Actually, the definition of this ID to be a uint32_t is not ideal and breaks layering.
-//However, it is difficult to keep layering, as the ID is stored in many places and sent around in
-//MessageQueueMessage.
-//Ideally, one would use the (current) object_id_t only, however, doing a lookup of queueIDs for every
-//call does not sound ideal.
-//In a first step, I'll circumvent the issue by not touching it, maybe in a second step.
-//This also influences Interface design (getCommandQueue) and some other issues..
-typedef uint32_t MessageQueueId_t;
 
 class MessageQueueSenderIF {
 public:
-	static const MessageQueueId_t NO_QUEUE = 0;
 
 	virtual ~MessageQueueSenderIF() {}
 
 	/**
-	 * Allows sending messages without actually "owing" a message queue.
+	 * Allows sending messages without actually "owning" a message queue.
 	 * Not sure whether this is actually a good idea.
-	 * Must be implemented by a subclass.
 	 */
 	static ReturnValue_t sendMessage(MessageQueueId_t sendTo,
-			MessageQueueMessage* message, MessageQueueId_t sentFrom =
-					MessageQueueSenderIF::NO_QUEUE, bool ignoreFault=false);
+			MessageQueueMessageIF* message,
+			MessageQueueId_t sentFrom = MessageQueueIF::NO_QUEUE,
+			bool ignoreFault = false);
 private:
 	MessageQueueSenderIF() {}
 };
 
 
-
 #endif /* FRAMEWORK_IPC_MESSAGEQUEUESENDERIF_H_ */
diff --git a/ipc/messageQueueDefinitions.h b/ipc/messageQueueDefinitions.h
new file mode 100644
index 000000000..d30e69849
--- /dev/null
+++ b/ipc/messageQueueDefinitions.h
@@ -0,0 +1,18 @@
+#ifndef FSFW_IPC_MESSAGEQUEUEDEFINITIONS_H_
+#define FSFW_IPC_MESSAGEQUEUEDEFINITIONS_H_
+
+#include <cstdint>
+
+/*
+ * TODO: Actually, the definition of this ID to be a uint32_t is not  ideal and
+ * breaks layering. However, it is difficult to keep layering, as the ID is
+ * stored in many places and sent around in MessageQueueMessage.
+ * Ideally, one would use the (current) object_id_t only, however, doing a
+ * lookup of queueIDs for every call does not sound ideal.
+ * In a first step, I'll circumvent the issue by not touching it,
+ * maybe in a second step. This also influences Interface design
+ * (getCommandQueue) and some other issues..
+ */
+using MessageQueueId_t = uint32_t;
+
+#endif /* FSFW_IPC_MESSAGEQUEUEDEFINITIONS_H_ */