Cloud-technology interplay simulator
In a recent project aiming to compare “distributed LTE enabled telco-cloud” solutions versus current best practice (i.e. Wi-Fi and Internet-based cloud), we developed a Pac-Man inspired simulator to show this comparison in a an easy to understand manner.
Theoretical advantages of LTE are the larger coverage area and the use of regulated spectrum. However, how and if the technical performance in responsiveness, security, reliability, network control and manageability actually address customer requirements for mission-critical cloud solutions in a superior way depends very much on the deployment scenarios. Hence, a holistic view in combination with clearly defined technical requirements is needed to make a real comparison.
To make this somewhat fuzzy problem more concrete and tangible we needed a scenario and a way to create understanding around these parameters and their combinational output. We needed to translate them into abstractions and give the possibility to free exploration.
As a simple demo scenario we imagined a remote controlled ‘work-site’ where you control the amount of workers (trucks or whatever) and the speed of them. During the sessions exploring narratives to wrap around this scenario - the1980’s arcade game Pac-Man came to mind. The idea was to use the game of Pac-Man as an abstract visualization of the scenario 'work-site', e.g. a factory or other automated industrial production facility.
After concluded that this idea was worth pursuing, we mapped the different technology parameters and suitable Pac-Man representations into the following basic scheme:
Since it is nearly impossible to foresee and predict the result of how the different technology parameters affect each other, the simulator relies on “emergent behavior” – which is high-level effects from low-level interactions. This is achieved by a set of interactive knobs to control the system parameters.
In the simulation you can manipulate three operational parameters and two resilience parameters. The first operational parameter is the number of equipment/workers the facility should have. This is represented as a one-to-one mapping to the number of Pac-Man(s) on the game board. The second one is the possibility to control the velocity of your workers - represented as visual feedback i.e. the animation duration it takes for the Pac-Man(s) to move from one spot to another. The third and final operational parameter is the wanted amount of bandwidth.
The combination of these three parameters affects the playback of the real-time game board. For example: low bandwidth and high number of workers affects the throughput thus affecting the latency. Latency is represented as an initial delay every time the Pac-Man is moving from one spot to another. High latency will render a very clear visual feedback of jerkiness in movement.
The two additional interactive parameters are Encryption and Interference cancellation. Turning these ON or OFF is represented in the number of problems (ghosts) on the game board. With the resilience parameters turned OFF, the Pac-Man(s) share the game board with double the amount of ghosts. If a Pac-Man collides with a ghost, it freezes up for a period of time. The duration of the freeze is also dependent on the resilience parameters.
Interaction with the parameters gives different results in worker movements. Since the goal for the Pac-Mans(s) is to collect pick-up’s to generate score (revenue) - smooth and swift movements are vital. The effect of latency and collisions with ghosts equals slow movement, jerkiness and freezes, which impacts the efficiency of collecting pick-ups and gives lower revenue.
Having the abstracted real-time view of two cloud-setups (game boards) next to each other - showing changes over time - makes it very easy to visually compare states, values and effects like: movement smoothness and revenue generation. In addition to the game boards we placed metric values of the parameters combinational output in the center of the UI. Basically represented as costs of different sorts, like: financial, power consumption and network load. As with the real-time view, the metric values are dynamically updated to reflect the result of the input parameters.
The comparison is a process of exploration and discovery. By adjusting the parameters one can instantly see the effect at the 'work-site', which gives a sense of how the technical parameters would interplay and affect a real world scenario.