Solvice Maps: Infrastructure and Deployment Guide

Infrastructure Overview

Solvice Maps runs on Google Cloud Platform (GCP) using a modern cloud-native architecture designed for high availability, scalability, and operational excellence. The infrastructure supports both real-time routing services and batch processing workloads with automatic scaling and comprehensive monitoring.

Cloud Architecture

Platform: Google Cloud Platform

Project Structure:

Primary Project: solver-285414
Primary Region: europe-west1 (Belgium)
Availability Zone: europe-west1-b
Secondary Regions: Available for multi-region deployment

Key GCP Services Used:

Compute: Google Compute Engine + Google Kubernetes Engine
Storage: Cloud Storage for large results and OSRM map data
Database: Cloud SQL (PostgreSQL) for request metadata
Messaging: Cloud Pub/Sub for event-driven processing
Networking: Global Load Balancer with Cloud CDN
Monitoring: Cloud Monitoring + Cloud Logging
Security: Cloud IAM + Secret Manager

Service Deployment Architecture

1. MapR Gateway Service (Primary API)

Deployment Platform: Google Cloud Run

Runtime: JVM 17 with Quarkus native compilation
Container: Distroless base image for security
Scaling: 0-100 instances with request-based auto-scaling
Cold Start: < 100ms with native compilation

Resource Configuration:

resources:
  limits:
    cpu: "2"
    memory: "4Gi"
  requests:
    cpu: "0.5"
    memory: "1Gi"

concurrency: 100
timeout: 300s

Environment Variables:

# Database Connection
DATABASE_URL=postgresql://user:pass@host:5432/mapr_gateway
DB_MAX_POOL_SIZE=20

# External Service Endpoints
OSRM_SERVICE_URL=https://osrm-europe.solvice.io
TOMTOM_API_KEY=${TOMTOM_API_KEY}
GOOGLE_MAPS_API_KEY=${GOOGLE_MAPS_API_KEY}

# Pub/Sub Configuration
PUBSUB_PROJECT_ID=solver-285414
PUBSUB_TABLE_TOPIC=mapr-table-requests
PUBSUB_RESPONSE_TOPIC=mapr-table-responses

# Storage Configuration
STORAGE_BUCKET=mapr-gateway-results
STORAGE_SIGNED_URL_DURATION=3600

# Authentication
JWT_SECRET=${JWT_SECRET}
JWT_ISSUER=solvice-maps

2. OSRM Service (Routing Engine)

Deployment Platform: Google Kubernetes Engine (GKE)

Cluster: osrm-cluster (3 nodes, n1-highmem-2)
Node Pool: Container-Optimized OS with SSD persistent disks
Scaling: Horizontal Pod Autoscaler with custom metrics

Kubernetes Deployment:

apiVersion: apps/v1
kind: Deployment
metadata:
  name: nodejs-mapr-europe-car
  namespace: default
spec:
  replicas: 2
  selector:
    matchLabels:
      app: nodejs-mapr-europe-car
  template:
    metadata:
      labels:
        app: nodejs-mapr-europe-car
    spec:
      initContainers:
      - name: map-downloader
        image: gcr.io/solver-285414/map-downloader:latest
        volumeMounts:
        - name: osrm-maps
          mountPath: /maps
        env:
        - name: MAPS_BUCKET
          value: "osrm-maps-europe"
        - name: MAP_REGION
          value: "europe"
      
      containers:
      - name: osrm-service
        image: gcr.io/solver-285414/nodejs-mapr:latest
        ports:
        - containerPort: 3000
        env:
        - name: OSRM_MAPS
          value: |
            [{
              "map": "europe",
              "vehicle": "car",
              "path": "/maps/europe-{{slice}}.osrm",
              "slices": [0,1,2,3,4,5,6,7,8,9,10,11,12],
              "mmap": true
            }]
        - name: PUBSUB_TABLE_SUBSCRIPTIONS
          value: |
            [{
              "id": "europe-car-subscription",
              "weight": 10,
              "maxMessages": 2
            }]
        volumeMounts:
        - name: osrm-maps
          mountPath: /maps
          readOnly: true
        resources:
          requests:
            memory: "8Gi"
            cpu: "2"
          limits:
            memory: "12Gi" 
            cpu: "4"
        livenessProbe:
          httpGet:
            path: /v1/health
            port: 3000
          initialDelaySeconds: 30
          periodSeconds: 10
        readinessProbe:
          httpGet:
            path: /v1/health/ready
            port: 3000
          initialDelaySeconds: 5
          periodSeconds: 5
          
      volumes:
      - name: osrm-maps
        persistentVolumeClaim:
          claimName: osrm-maps-pvc

Auto-Scaling Configuration:

apiVersion: autoscaling/v2
kind: HorizontalPodAutoscaler
metadata:
  name: osrm-service-hpa
spec:
  scaleTargetRef:
    apiVersion: apps/v1
    kind: Deployment
    name: nodejs-mapr-europe-car
  minReplicas: 2
  maxReplicas: 10
  metrics:
  - type: Resource
    resource:
      name: cpu
      target:
        type: Utilization
        averageUtilization: 70
  - type: Resource
    resource:
      name: memory
      target:
        type: Utilization
        averageUtilization: 80
  - type: External
    external:
      metric:
        name: pubsub.googleapis.com/subscription/num_undelivered_messages
        selector:
          matchLabels:
            resource.labels.subscription_id: "europe-car-subscription"
      target:
        type: AverageValue
        averageValue: "30"

Infrastructure as Code (Terraform)

Terraform Configuration Structure

terraform/
├── backend.tf          # Remote state configuration
├── provider.tf         # GCP provider configuration
├── variables.tf        # Input variables
├── container.tf        # GCE + Container configuration
├── instance_template.tf # VM instance template
├── instance_group.tf   # Managed instance group
├── autoscaler.tf       # Auto-scaling configuration
├── loadbalancer.tf     # Global load balancer
├── health_check.tf     # Health check configuration
├── bucket.tf           # Cloud Storage buckets
└── pubsub.tf          # Pub/Sub topics and subscriptions

Key Terraform Resources

Instance Template:

resource "google_compute_instance_template" "osrm_template" {
  name_prefix  = "osrm-template-"
  description  = "Template for OSRM service instances"
  machine_type = var.machine_type # n1-highmem-2

  disk {
    source_image = "cos-cloud/cos-stable"
    disk_type    = "pd-ssd"
    disk_size_gb = 280
    auto_delete  = true
    boot         = true
  }

  disk {
    source_image = var.data_disk_image # Custom image with OSRM data
    disk_type    = "pd-ssd"
    disk_size_gb = 1400
    auto_delete  = false
    boot         = false
  }

  network_interface {
    network = "default"
    access_config {
      nat_ip = null # Ephemeral IP
    }
  }

  metadata = {
    "gce-container-declaration" = module.gce-container.metadata_value
    "google-logging-enabled"    = "true"
    "enable-guest-attributes"   = "TRUE"
  }

  service_account {
    email  = var.service_account_email
    scopes = ["https://www.googleapis.com/auth/cloud-platform"]
  }

  tags = ["http-server", "https-server"]

  lifecycle {
    create_before_destroy = true
  }
}

Global Load Balancer:

resource "google_compute_global_forwarding_rule" "default" {
  name       = "osrm-global-forwarding-rule"
  target     = google_compute_target_http_proxy.default.id
  port_range = "80"
  ip_address = google_compute_global_address.default.address
}

resource "google_compute_target_http_proxy" "default" {
  name    = "osrm-target-proxy"
  url_map = google_compute_url_map.default.id
}

resource "google_compute_url_map" "default" {
  name            = "osrm-url-map"
  default_service = google_compute_backend_service.default.id
}

resource "google_compute_backend_service" "default" {
  name                  = "osrm-backend-service"
  protocol              = "HTTP"
  timeout_sec           = 30
  enable_cdn           = true
  load_balancing_scheme = "EXTERNAL"

  backend {
    group           = google_compute_instance_group_manager.default.instance_group
    balancing_mode  = "UTILIZATION"
    max_utilization = 0.8
  }

  health_checks = [google_compute_health_check.default.id]
}

Pub/Sub Configuration:

# Dynamic topic creation from JSON configuration
locals {
  pubsub_config = jsondecode(var.pubsub_subscriptions_json)
}

resource "google_pubsub_topic" "table_topics" {
  for_each = { for sub in local.pubsub_config : sub.id => sub }
  name     = "mapr-table-${each.value.id}"
  
  message_retention_duration = "604800s" # 7 days
}

resource "google_pubsub_topic" "dead_letter_topics" {
  for_each = { for sub in local.pubsub_config : sub.id => sub }
  name     = "mapr-table-${each.value.id}-dead-letter"
}

resource "google_pubsub_subscription" "table_subscriptions" {
  for_each = { for sub in local.pubsub_config : sub.id => sub }
  name     = "mapr-table-${each.value.id}-subscription"
  topic    = google_pubsub_topic.table_topics[each.key].name

  ack_deadline_seconds       = 600
  message_retention_duration = "604800s"
  retain_acked_messages      = false

  retry_policy {
    minimum_backoff = "10s"
    maximum_backoff = "600s"
  }

  dead_letter_policy {
    dead_letter_topic     = google_pubsub_topic.dead_letter_topics[each.key].id
    max_delivery_attempts = 5
  }
}

Deployment Processes

1. CI/CD Pipeline (GitLab CI)

Pipeline Structure:

stages:
  - validate
  - test
  - build
  - deploy-staging
  - integration-test
  - deploy-production

variables:
  DOCKER_DRIVER: overlay2
  DOCKER_TLS_CERTDIR: "/certs"

# Terraform Validation
validate:
  stage: validate
  image: hashicorp/terraform:1.9
  script:
    - cd terraform
    - terraform init -backend=false
    - terraform validate
    - terraform fmt -check

# Application Testing
test:
  stage: test
  image: node:22
  script:
    - cd osrm-service
    - npm ci
    - npm run test:unit
    - npm run test:integration

# Container Build
build:
  stage: build
  image: docker:24
  services:
    - docker:24-dind
  script:
    - docker build -t $CI_REGISTRY_IMAGE:$CI_COMMIT_SHA .
    - docker push $CI_REGISTRY_IMAGE:$CI_COMMIT_SHA

# Staging Deployment
deploy-staging:
  stage: deploy-staging
  image: google/cloud-sdk:alpine
  script:
    - gcloud auth activate-service-account --key-file $GOOGLE_APPLICATION_CREDENTIALS
    - gcloud config set project $GCP_PROJECT_ID
    - cd terraform
    - terraform init
    - terraform workspace select staging
    - terraform plan -var="image_tag=$CI_COMMIT_SHA"
    - terraform apply -auto-approve -var="image_tag=$CI_COMMIT_SHA"
  environment:
    name: staging
    url: https://staging-api.solvice.io

# Production Deployment (Manual)
deploy-production:
  stage: deploy-production
  image: google/cloud-sdk:alpine
  script:
    - gcloud auth activate-service-account --key-file $GOOGLE_APPLICATION_CREDENTIALS
    - gcloud config set project $GCP_PROJECT_ID
    - cd terraform
    - terraform init
    - terraform workspace select production
    - terraform plan -var="image_tag=$CI_COMMIT_SHA"
    - terraform apply -auto-approve -var="image_tag=$CI_COMMIT_SHA"
  environment:
    name: production
    url: https://routing.solvice.io
  when: manual
  only:
    - main

2. Zero-Downtime Deployment Strategy

Rolling Update Process:

Health Check: Ensure all current instances are healthy
New Instance Launch: Launch new instances with updated configuration
Health Validation: Wait for new instances to pass health checks
Traffic Migration: Gradually shift traffic to new instances
Old Instance Termination: Terminate old instances after validation
Rollback Plan: Automated rollback if health checks fail

Blue-Green Deployment for Critical Updates:

#!/bin/bash
# Blue-Green deployment script

# Deploy to blue environment
terraform workspace select blue
terraform apply -var="image_tag=$NEW_VERSION"

# Run health checks
./scripts/health-check.sh blue

# Switch traffic to blue
gcloud compute url-maps set-default-service $URL_MAP \
  --default-service=$BLUE_BACKEND_SERVICE

# Monitor for 10 minutes
sleep 600

# If successful, cleanup green environment
if [ $? -eq 0 ]; then
  terraform workspace select green
  terraform destroy -auto-approve
  echo "Deployment successful"
else
  # Rollback to green
  gcloud compute url-maps set-default-service $URL_MAP \
    --default-service=$GREEN_BACKEND_SERVICE
  echo "Deployment failed, rolled back"
  exit 1
fi

3. Map Data Deployment

OSRM Map Update Process:

#!/bin/bash
# Map data update script

# Build new map data
./build-osrm-maps.sh $REGION $VERSION

# Create disk image
gcloud compute images create osrm-$REGION-$VERSION \
  --source-disk=osrm-build-disk \
  --source-disk-zone=europe-west1-b

# Update Terraform variable
export TF_VAR_data_disk_image="osrm-$REGION-$VERSION"

# Deploy with rolling update
terraform plan -var="data_disk_image=$TF_VAR_data_disk_image"
terraform apply -auto-approve

Monitoring and Alerting

1. Infrastructure Monitoring

Cloud Monitoring Metrics:

# Custom metric for OSRM request latency
- name: "osrm/request_duration_seconds"
  description: "OSRM request processing time"
  type: "histogram"
  labels: ["method", "status", "region"]

# Custom metric for queue depth
- name: "pubsub/queue_depth"
  description: "Number of undelivered messages"
  type: "gauge"
  labels: ["subscription", "topic"]

# Infrastructure metrics
- name: "compute/cpu_utilization"
- name: "compute/memory_utilization"
- name: "compute/disk_utilization"

Alerting Policies:

alertPolicy:
  displayName: "High Response Time"
  conditions:
    - displayName: "Response time > 100ms"
      conditionThreshold:
        threshold: 0.1
        comparison: COMPARISON_GREATER_THAN
        metric: "osrm/request_duration_seconds"
        aggregations:
          - alignmentPeriod: "300s"
            perSeriesAligner: ALIGN_RATE
  notificationChannels:
    - "projects/solver-285414/notificationChannels/slack-alerts"
    - "projects/solver-285414/notificationChannels/pager-duty"

2. Application-Level Monitoring

Health Check Endpoints:

// Comprehensive health checks
@Get('/health')
async getHealth(): Promise<HealthStatus> {
  return {
    status: 'healthy',
    timestamp: new Date().toISOString(),
    services: {
      database: await this.checkDatabase(),
      osrm: await this.checkOSRM(),
      pubsub: await this.checkPubSub(),
      storage: await this.checkStorage()
    },
    metrics: {
      activeConnections: this.getActiveConnections(),
      queueDepth: await this.getQueueDepth(),
      memoryUsage: process.memoryUsage()
    }
  };
}

Performance Metrics:

// Custom metrics collection
@Histogram('request_duration_seconds', ['method', 'status'])
private requestDuration: Histogram;

@Counter('requests_total', ['method', 'status'])
private requestsTotal: Counter;

@Gauge('active_requests', [])
private activeRequests: Gauge;

Security Configuration

1. Network Security

VPC Configuration:

resource "google_compute_network" "solvice_vpc" {
  name                    = "solvice-maps-vpc"
  auto_create_subnetworks = false
}

resource "google_compute_subnetwork" "private_subnet" {
  name          = "private-subnet"
  ip_cidr_range = "10.0.1.0/24"
  region        = "europe-west1"
  network       = google_compute_network.solvice_vpc.id
  
  private_ip_google_access = true
}

resource "google_compute_firewall" "allow_internal" {
  name    = "allow-internal"
  network = google_compute_network.solvice_vpc.name

  allow {
    protocol = "tcp"
    ports    = ["80", "443", "3000"]
  }

  source_ranges = ["10.0.0.0/8"]
}

SSL/TLS Configuration:

resource "google_compute_managed_ssl_certificate" "default" {
  name = "solvice-maps-ssl-cert"

  managed {
    domains = [
      "routing.solvice.io",
      "api.solvice.io"
    ]
  }
}

resource "google_compute_target_https_proxy" "default" {
  name             = "solvice-https-proxy"
  url_map          = google_compute_url_map.default.id
  ssl_certificates = [google_compute_managed_ssl_certificate.default.id]
}

2. IAM and Access Control

Service Account Configuration:

resource "google_service_account" "osrm_service_account" {
  account_id   = "osrm-service"
  display_name = "OSRM Service Account"
  description  = "Service account for OSRM compute instances"
}

resource "google_project_iam_member" "osrm_storage_access" {
  project = var.project_id
  role    = "roles/storage.objectViewer"
  member  = "serviceAccount:${google_service_account.osrm_service_account.email}"
}

resource "google_project_iam_member" "osrm_pubsub_access" {
  project = var.project_id
  role    = "roles/pubsub.subscriber"
  member  = "serviceAccount:${google_service_account.osrm_service_account.email}"
}

Secret Management:

resource "google_secret_manager_secret" "api_keys" {
  secret_id = "external-api-keys"
  
  replication {
    user_managed {
      replicas {
        location = "europe-west1"
      }
    }
  }
}

resource "google_secret_manager_secret_version" "api_keys_version" {
  secret      = google_secret_manager_secret.api_keys.id
  secret_data = jsonencode({
    tomtom_api_key = var.tomtom_api_key
    google_maps_api_key = var.google_maps_api_key
  })
}

Disaster Recovery and Backup

1. Data Backup Strategy

Database Backups:

# Automated PostgreSQL backups
gcloud sql backups create \
  --instance=mapr-gateway-db \
  --description="Daily automated backup $(date +%Y-%m-%d)"

# Point-in-time recovery enabled
gcloud sql instances patch mapr-gateway-db \
  --backup-start-time=02:00 \
  --enable-bin-log

Configuration Backups:

# Terraform state backup
gsutil cp gs://terraform-state-bucket/terraform.tfstate \
  gs://disaster-recovery-bucket/terraform-$(date +%Y%m%d).tfstate

# Container images backup
gcloud container images list-tags gcr.io/solver-285414/nodejs-mapr \
  --limit=10 --format='get(digest)' | \
  xargs -I {} gcloud container images add-tag \
    gcr.io/solver-285414/nodejs-mapr@{} \
    gcr.io/backup-project/nodejs-mapr:backup-$(date +%Y%m%d)

2. Multi-Region Deployment

Regional Failover Configuration:

# Primary region: europe-west1
# Secondary region: us-central1

resource "google_compute_instance_group_manager" "osrm_primary" {
  name     = "osrm-primary"
  location = "europe-west1-b"
  # ... primary configuration
}

resource "google_compute_instance_group_manager" "osrm_secondary" {
  name     = "osrm-secondary" 
  location = "us-central1-b"
  # ... secondary configuration (standby)
}

resource "google_compute_health_check" "regional_failover" {
  name = "regional-failover-check"
  
  http_health_check {
    port         = 80
    request_path = "/health"
  }
  
  check_interval_sec  = 10
  timeout_sec         = 5
  healthy_threshold   = 2
  unhealthy_threshold = 3
}

Cost Optimization

1. Resource Optimization

Preemptible Instances:

resource "google_compute_instance_template" "preemptible_template" {
  name = "osrm-preemptible-template"
  
  scheduling {
    preemptible = true
    automatic_restart = false
    on_host_maintenance = "TERMINATE"
  }
  
  # Use preemptible instances for batch processing
  machine_type = "n1-highmem-2"
}

Auto-Scaling Configuration:

resource "google_compute_autoscaler" "osrm_autoscaler" {
  name   = "osrm-autoscaler"
  target = google_compute_instance_group_manager.default.id

  autoscaling_policy {
    max_replicas    = 10
    min_replicas    = 1  # Scale to zero during off-hours
    cooldown_period = 300

    cpu_utilization {
      target = 0.7
    }

    scaling_schedules {
      name                  = "scale-down-nights"
      description           = "Scale down during off-hours"
      schedule              = "0 22 * * *"  # 10 PM
      time_zone             = "Europe/Brussels"
      min_required_replicas = 0
      duration_sec          = 28800  # 8 hours
    }
  }
}

This infrastructure provides a robust, scalable, and cost-effective foundation for the Solvice Maps platform, with comprehensive monitoring, security, and disaster recovery capabilities.

Get Started

Directions

Matrix

Cube

Infrastructure deployment

Solvice Maps: Infrastructure and Deployment Guide

Infrastructure Overview

Cloud Architecture

Platform: Google Cloud Platform

Service Deployment Architecture

1. MapR Gateway Service (Primary API)

2. OSRM Service (Routing Engine)

Infrastructure as Code (Terraform)

Terraform Configuration Structure

Key Terraform Resources

Deployment Processes

1. CI/CD Pipeline (GitLab CI)

2. Zero-Downtime Deployment Strategy

3. Map Data Deployment

Monitoring and Alerting

1. Infrastructure Monitoring

2. Application-Level Monitoring

Security Configuration

1. Network Security

2. IAM and Access Control

Disaster Recovery and Backup

1. Data Backup Strategy

2. Multi-Region Deployment

Cost Optimization

1. Resource Optimization

Get Started

Directions

Matrix

Cube

​Solvice Maps: Infrastructure and Deployment Guide

​Infrastructure Overview

​Cloud Architecture

​Platform: Google Cloud Platform

​Service Deployment Architecture

​1. MapR Gateway Service (Primary API)

​2. OSRM Service (Routing Engine)

​Infrastructure as Code (Terraform)

​Terraform Configuration Structure

​Key Terraform Resources

​Deployment Processes

​1. CI/CD Pipeline (GitLab CI)

​2. Zero-Downtime Deployment Strategy

​3. Map Data Deployment

​Monitoring and Alerting

​1. Infrastructure Monitoring

​2. Application-Level Monitoring

​Security Configuration

​1. Network Security

​2. IAM and Access Control

​Disaster Recovery and Backup

​1. Data Backup Strategy

​2. Multi-Region Deployment

​Cost Optimization

​1. Resource Optimization

Solvice Maps: Infrastructure and Deployment Guide

Infrastructure Overview

Cloud Architecture

Platform: Google Cloud Platform

Service Deployment Architecture

1. MapR Gateway Service (Primary API)

2. OSRM Service (Routing Engine)

Infrastructure as Code (Terraform)

Terraform Configuration Structure

Key Terraform Resources

Deployment Processes

1. CI/CD Pipeline (GitLab CI)

2. Zero-Downtime Deployment Strategy

3. Map Data Deployment

Monitoring and Alerting

1. Infrastructure Monitoring

2. Application-Level Monitoring

Security Configuration

1. Network Security

2. IAM and Access Control

Disaster Recovery and Backup

1. Data Backup Strategy

2. Multi-Region Deployment

Cost Optimization

1. Resource Optimization