Some of the previously in Section7defined sequences are tested and discussed in the subsequent Sections. Both scenarios are simulated under the following conditions.
• The same message-sequences are used in the monolithic and the distributed scenario.
• The messages are sent directly to the receiver, if the proposed platform is used.
• The messages are sent via the central platform, if a monolithic platform is used.
• The Gateway, Management and Registry are combined to one component in the mono-lithic approach and therefore no messages between those three components are mea-sured.
• A message is sent for every request. The messages are not buffered and not sent together or not sent in smaller groups.
• Initial message rhythm.
0s - 10s set-up time.
10s - 30s 10 initial messages per second were sent with equally temporal distance.
30s - 70s 50 initial messages per second were sent with equally temporal distance.
70s - 110s 20 initial messages per second were sent with equally temporal distance.
8.2.1. Scenario Object Engine Interaction
The communication between a User and an Object Engine is chosen as a relevant scenario. In this scenario the Object Engine is already registered with the platform. The registration was done with sequence7.3.2. The interaction that is measured, is done with messages described in sequence7.3.5.
The messageinvokeObjectMethod A.2.22is the initial message for the sequence. The messages are sent in the rhythms, that are described before. The address and the information of the Object Engine are prescribed and in order to that a look-up was not necessary.
In the diagram, that is shown in Figure8.3, the results of the measurements, that were done in scenario of the Object Engine interaction, are presented.
The data was collected as an OMNeT++ vector, that was exported to a python array. The data in the array was evaluated and reprocessed. Afterwards the data was illustrated in the diagram.
The horizontal axis is defined as the time in seconds and the vertical axis is defined as messages per second. In the distributed line graph, four Sections can be seen. First, from zero seconds to ten seconds, second from ten seconds up to 30 seconds, third from 30 seconds up to 70 seconds and fourth from 70 seconds up to 110 seconds. In the subsequent description the amounts of messages per second, that are indicated by the distributed line graph, are summarized.
Section one nearly zero messages per second Section two approximately 20 messages per second Section three approximately 100 messages per second Section four approximately 40 messages per second
In the monolithic line graph the same four time-sections can be seen. In the subsequent description the amounts of messages per second, that are indicated by the monolithic line graph, are summarized.
Section one nearly zero messages per second Section two approximately 40 messages per second Section three approximately 200 messages per second Section four approximately 80 messages per second
Like it is mentioned before, the same four time-sections are shown by both line graphs. It can be seen, that the amount of messages per second is twice as high with the monolithic platform. The amount of messages with the monolithic platform is factor two higher as in the distributed approach.
The reason for the difference can be seen in the monolithic platform. With the monolithic platform all messages are sent not only between the User and the Object Engine, but the User to the central instance to the Object Engine. That behavior can be seen in Figure8.4. With the distributed platform only path A (Figure8.4) is used. In the monolithic platform the paths B and C (Figure8.4) are used. Paths in this scenario are bidirectional connections between two components, that can be used to send a message.
0 10 20 30 40 50 60 70 80 90 100 110 time [s]
0 25 50 75 100 125 150 175 200
messages per second
Monolithic Distributed
Figure 8.3.: Scenario Object Engine interaction measurements
Figure 8.4.: Path usage in monolithic and distributed scenarios 8.2.2. Scenario Sub-Registry Registration
The registration as a sub-Registry is chosen as another relevant scenario.
The interaction that is measured, is done with messages described in sequence7.3.3. The messageregisterAsSubRegistryA.2.40is chosen to be the initial message for the sequence.
The messages are sent in the rhythms, that are described before. The address and the informa-tion of the Object Engine are prescribed and because of this a look-up was not necessary.
In the diagram, that is shown in Figure8.5, the results of the measurements, that were done in the sub-Registry registration scenario, are described. The data was collected as an OMNeT++
vector, that was exported to a python array. The data in the array was evaluated and reprocessed.
Afterwards the data was illustrated in the diagram. The horizontal axis is defined as the time in seconds and the vertical axis is defined as messages per second. In the distributed line graph four Sections can be seen. First, from zero seconds to ten seconds, second from ten seconds up to 30 seconds, third from 30 seconds up to 70 seconds and fourth from 70 seconds up to 110 seconds. In the subsequent description the amounts of messages per second, that are indicated by the distributed line graph, are summarized.
Section one nearly zero messages per second Section two approximately 60 messages per second Section three approximately 300 messages per second Section four approximately 120 messages per second
In the monolithic line graph the same four time-sections can be seen. In the subsequent description the amounts of messages per second, that are indicated by the monolithic line graph, are summarized.
Section one nearly zero messages per second Section two approximately 40 messages per second Section three approximately 200 messages per second Section four approximately 80 messages per second
Like it is mentioned before, the same four time-sections are shown by both line graphs. It can be seen, that the amount of messages per second is only two thirds with the monolithic platform. The amount of messages with the distributed platform is factor one and a half higher then with the monolithic platform.
The amount of messages that was measured with the distributed approach is higher. This is the result of the distribution of the the components Management and Registry. With the monolithic platform those messages are part of the internal communication and are therefore not taken into account.
0 10 20 30 40 50 60 70 80 90 100 110 time [s]
0 50 100 150 200 250 300
messages per second
Monolithic Distributed
Figure 8.5.: Scenario Object Engine registration measurements