Plan your deployment

This section describes the prerequisites, supported Regions, cost, security, and quota considerations before deploying the guidance.

Prerequisites

Before deploying the guidance, ensure you have the following prerequisites:

Software requirements

Required software:

AWS account requirements

Account setup:

Required IAM permissions:

The IAM user or role deploying the guidance needs permissions to create and manage:

We recommend using the AdministratorAccess managed policy for initial deployment, then creating a custom policy with least-privilege permissions for production deployments.

Network requirements

VPC considerations:

The solution creates a new VPC by default with the following configuration:

Existing VPC:

To use an existing VPC, set the VPC_ID environment variable before deployment. The VPC must have:

Supported AWS Regions

For the most current availability of AWS services by Region, see the AWS Regional Services List.

Guidance for Connected Mobility on AWS is supported in the following AWS Regions:

Cost

You are responsible for the cost of the AWS services used while running this solution. Prices are subject to change. For full details, see the pricing webpage for each AWS service used in this solution.

We recommend creating a budget through AWS Cost Explorer to help manage costs.

Cost model assumptions

The cost estimates below are based on the following telemetry profile per vehicle:

All estimates use US East (N. Virginia) pricing as of March 2026.

Cost by fleet size

Cost per vehicle per month:

The per-vehicle cost drops dramatically as fleet size increases because the fixed infrastructure costs (MSK cluster, NAT Gateway, ElastiCache) are amortized across more vehicles. The breakpoint for cost efficiency is around 5,000–10,000 vehicles, where per-vehicle cost drops below $0.15/month.

Detailed cost breakdown

Amazon MSK (message streaming)

MSK is the largest cost driver (40-50% of total). Cost scales with broker count and instance size.

Scaling trigger: Upgrade when average broker CPU exceeds 60% or when consumer lag exceeds 30 seconds.

Storage: 100 GB per broker included. Telemetry retention is 7 days. At 1,000 vehicles, daily ingest is ~7 GB, so 100 GB per broker provides comfortable headroom.

Amazon Managed Service for Apache Flink (stream processing)

Flink is the second largest cost driver (25-35% of total). Cost scales with KPU count. Each KPU provides 1 vCPU and 4 GB memory at $0.15/hour ($108/month).

The solution runs 7-10 Flink applications. In development, each application uses 1 KPU. In production, data-path processors (SimulatorPreprocessor, EventDrivenTelemetryProcessor, TripProcessor) may need 2-4 KPUs each.

Scaling trigger: Add KPUs when millisBehindLatest exceeds 5,000ms or when checkpoint duration exceeds 50% of the checkpoint interval.

Infrastructure (VPC, ElastiCache, NAT Gateway)

These are mostly fixed costs that do not scale linearly with fleet size.

ElastiCache scaling: The Redis node must hold the Last Known State for all active vehicles. Each vehicle uses ~5 KB in Redis (signals hash + timestamps hash + meta hash + stream). At 10,000 vehicles, that is ~50 MB — well within a cache.t3.micro (0.5 GB). Upgrade to cache.t3.small at 50,000+ vehicles or if geospatial query latency exceeds 5ms.

Application layer (Lambda, API Gateway, IoT Core, DynamoDB, S3)

These costs scale with usage but remain a small percentage of total cost.

IoT Core message cost detail: At 1,000 vehicles × 3,600 messages/day = 3.6M messages/day = 108M messages/month. IoT Core charges $1.00 per million messages (first 1B), so 108M × $1.00/M = $108. However, messages are metered in 5 KB increments, and compressed telemetry is ~2 KB, so each message counts as 1 unit. For 100 vehicles, this drops to $10.80/month.

DynamoDB cost detail: On-demand pricing is $1.25 per million write request units and $0.25 per million read request units. The stateful TripProcessor design reduces writes by 80% compared to a stateless approach (see Trip lifecycle).

Cost breakpoints and optimization

The cost curve has three distinct regions:

Under 500 vehicles (~$265/month): Fixed infrastructure dominates. MSK and Flink minimum billing account for 66% of cost. Per-vehicle cost is high ($0.53–$2.65) but total cost is low. Use kafka.t3.small brokers and 1 KPU per Flink app.

500–10,000 vehicles (~$400–$1,075/month): The sweet spot. Infrastructure costs are amortized, and usage-based costs (IoT Core, DynamoDB) are still modest. Per-vehicle cost drops to $0.11–$0.40. This is where the architecture is most cost-efficient relative to capability.

Over 10,000 vehicles (~$1,075+/month): Usage-based costs begin to dominate. MSK and Flink need to scale horizontally. Per-vehicle cost flattens at ~$0.05. At this scale, consider:

Cost optimization strategies

Development environment:

Production optimizations:

Security

Data protection

Encryption at rest:

Encryption in transit:

Identity and access management

Authentication:

Authorization:

Network security

VPC isolation:

API security:

Monitoring and logging

CloudWatch Logs:

CloudTrail:

Compliance

This solution uses AWS services that support various compliance programs:

For the most current compliance information, see AWS Services in Scope by Compliance Program.

Quotas

Service quotas, also referred to as limits, are the maximum number of service resources or operations for your AWS account.

Service quotas to verify

Before deploying the guidance, verify you have sufficient quotas for the following services:

Amazon MSK:

Amazon Kinesis Data Analytics:

AWS IoT Core:

Amazon DynamoDB:

AWS Lambda:

Amazon VPC:

Amazon ElastiCache:

Requesting quota increases

If you need to increase service quotas:

Most quota increases are processed within 24-48 hours.

Deployment sizing

Small fleet (100-1,000 vehicles)

Recommended configuration:

Expected cost: ~$250-300/month

Telemetry capacity:

Medium fleet (1,000-10,000 vehicles)

Recommended configuration:

Expected cost: ~$410-600/month

Telemetry capacity:

Large fleet (10,000+ vehicles)

Recommended configuration:

Expected cost: ~$1,200-2,000/month

Telemetry capacity:

Performance considerations

Message throughput:

Latency targets:

Scaling triggers:

Disaster recovery

Backup strategy

Automated backups:

Manual backups:

Recovery time objectives

RTO (Recovery Time Objective):

RPO (Recovery Point Objective):

Multi-region considerations

For high availability across regions: