Design principles

In addition to the general Well-Architected cost optimization design principles, there are some design principles specific for cost optimization for connected mobility:

Manage cost and business value tradeoffs: Managing tradeoffs between cost and other factors in connected mobility requires a careful balance between short-term costs, such as installation, setup and subscription fees, and long-term costs, such as data storage, management and infrastructure upgrades cost, cost-benefit analysis, and risk management. For example, if you want to optimize for speed to market with low cost?

Choosing a solution solely based on cost without considering its business value can result in increased operational cost overtime. Factors such as functionality, scalability, security, speed of innovation, and going to market should be carefully evaluated along with cost to select the right services/solutions. For example, a logistics company wants to implement a connected mobility solution to track packages and optimize delivery routes. Implementing a flexible architecture that can scale to accommodate growth and demand may come with an increased short-term cost, however, it also comes with an increased business value creation which will result in cost savings in long-term.

Design the system to filter relevant data: Connected mobility applications could generate 10 exabytes of data per month. That's why it is important to identify the data to be collected throughout your vehicle's fleets based on the corresponding business use-case or security/compliance requirements. Identify opportunities to stop collecting unnecessary data and consider collecting data targeted on business case and moving to event-based collection instead of interval whenever possible.

Evaluate where to process data: Data transfer between vehicle and cloud backend incur cost. Evaluation based on the use case through several dimensions:

• The network cost private or public may have different cost implications. In the context of connected vehicles, a private network refers to a dedicated and secure communication network established specifically for the vehicles and related systems within a closed environment. This network is isolated from public networks and is designed to provide a controlled and reliable communication infrastructure for connected vehicle operations.

  For example, imagine a fleet of autonomous delivery vehicles operated by a logistics company. To ensure seamless and secure communication between these vehicles, as well as with the central control system, the company sets up a private network. This network could consist of dedicated cellular connections, Wi-Fi hotspots, or even a custom-built communication infrastructure.

• The latency requirements to process the data. Latency refers to the delay between when data is generated or transmitted and when it is received and processed by the relevant systems. Here are two examples of latency requirements for processing data in connected vehicles:

Emergency collision avoidance:

Data source: Sensors on a vehicle detect an imminent collision with another vehicle.

Processing needs: The data from these sensors must be processed immediately to assess the risk and decide on an appropriate action (for example, applying brakes, or changing lanes).

Latency tolerance: In this scenario, extremely low latency is critical. A delay of even a few milliseconds could be the difference between a successful collision avoidance and an accident.

Traffic signal optimization:

Data source: The connected vehicle is equipped with technology to communicate with traffic signals in real-time.

Processing needs: The vehicle's system sends data to the traffic signal control system to request a green light or to adjust the signal timing for optimal traffic flow.

Latency tolerance: While not as critical as collision avoidance, low latency is still important. Delays should be minimized to ensure smooth traffic flow and reduce unnecessary stops.

In both cases, low latency is essential for the effective operation of connected vehicle systems. It allows for timely decision-making and actions, enhancing safety and efficiency on the road.

• The compute capacity requirements where processing the data makes more sense, at edge in the vehicle or in the AWS Cloud. In this scenario, edge processing (in-vehicle) is critical for immediate decision-making and ensuring passenger safety. It allows each vehicle to operate autonomously, reacting swiftly to its environment.

On the other hand, cloud processing is beneficial for higher-level, aggregate analysis. It can be used to optimize routes across the entire fleet, predict traffic patterns, and perform long-term planning. Overall, a combination of edge and cloud processing can offer the best of both worlds, allowing for real-time decision-making at the vehicle level while also leveraging the cloud's computational power for broader fleet optimization and analytics.