- What are key similarities and differences between GCP, AWS, and Azure in terms of core services (compute, storage, networking)?
I made a table to compare the actual products side by side and cut the marketing fat out of the analysis.
| Category | Azure | AWS | GCP |
| Compute | Virtual Machines (VMs) via Azure Virtual Machines | EC2 instances (Elastic Compute Cloud) | Google Compute Engine (GCE) |
| Azure Functions (Serverless) | AWS Lambda (Serverless) | Cloud Functions (Serverless) | |
| Azure Kubernetes Service (AKS) | Amazon Elastic Kubernetes Service (EKS) | Google Kubernetes Engine (GKE) | |
| Storage | Azure Blob Storage | Amazon S3 | Google Cloud Storage |
| Azure Disk Storage (SSD, HDD options) | Amazon EBS (Elastic Block Store) | Google Persistent Disks | |
| Azure Archive Storage (cold tier) | Amazon Glacier | Coldline and Archive Storage | |
| Networking | Azure Virtual Network (VNet) | Amazon VPC (Virtual Private Cloud) | Google Virtual Private Cloud (VPC) |
| Azure ExpressRoute (private connection) | AWS Direct Connect | Google Cloud Interconnect | |
| Load Balancer, Traffic Manager, Application Gateway | Elastic Load Balancing (ELB), Route 53 | Google Cloud Load Balancing | |
| Database | Azure SQL Database (PaaS), Cosmos DB (NoSQL) | Amazon RDS (Relational), DynamoDB (NoSQL) | Cloud SQL (Relational), Firestore (NoSQL) |
| Azure Database for MySQL, PostgreSQL | Aurora (MySQL/PostgreSQL-compatible), Redshift (Data Warehouse) | Cloud Spanner (Globally distributed SQL DB), BigQuery (Data Warehouse) | |
| AI/ML Services | Azure Machine Learning, Cognitive Services | Amazon SageMaker, AWS AI/ML Services | AI Platform, Vertex AI |
| Azure OpenAI Service | AWS Deep Learning AMIs | Google DeepMind, AutoML | |
| Global Network | 60+ regions | 100+ availability zones in 32 regions | 35 regions, 106 zones |
| Hybrid Solutions | Azure Arc (for hybrid and multi-cloud deployments) | AWS Outposts (extend AWS on-premises) | Anthos (multi-cloud and hybrid management) |
| Containers | Azure Container Instances, Azure Kubernetes Service (AKS) | Amazon ECS (Elastic Container Service), EKS | Google Kubernetes Engine (GKE), Cloud Run |
| Security | Azure Security Center, Azure Active Directory | AWS Identity and Access Management (IAM), AWS Shield | Google Cloud IAM, Google Security Command Center |
| Monitoring | Azure Monitor, Application Insights | AWS CloudWatch, X-Ray | Google Cloud Monitoring, Google Cloud Operations Suite |
| Pricing Models | Pay-as-you-go, Reserved VMs, Spot pricing | On-demand, Reserved Instances, Spot Instances | Pay-as-you-go, Sustained Use Discounts, Committed Use Contracts |
| DevOps Tools | Azure DevOps, GitHub (owned by Microsoft), Azure Pipelines | AWS CodePipeline, CodeCommit | Google Cloud Build, Cloud Source Repositories |
| Big Data | Azure HDInsight, Azure Synapse Analytics | Amazon EMR (Elastic MapReduce), Redshift | Google BigQuery, Cloud Dataflow |
| IoT Solutions | Azure IoT Hub, Azure Digital Twins | AWS IoT Core, AWS Greengrass | Google Cloud IoT Core, Edge TPU |
| Free Tier | 12-month free services, limited always free | 12-month free services, always free tier (750 hours/month EC2) | 12-month free services, always free (300 USD credit for new users) |
- Which platform do you think is more user-friendly for beginners, and why?
I believe they are all predatory but Amazon has the least ulterior motive and greatest value proposition of the three options.
- How do you see the competition between AWS, GCP, and Azure evolving in the next 5 years?
Historically, Amazon actually sells goods and services and is customer service focused, whereas Microsoft and Google are both focused on virtual products and provide little, if any, service and support. Amazon offers training and education for free, whereas Google and Microsoft have freemium and paid tiers to bait and switch customers like cloud addicts. Microsoft has a bigger share price, but it has historically had an inflated value due to government contracts and engineered dependency of financial technology and businesses on Microsoft products. Amazon is consistently valued and is more flexible and less ego driven than Microsoft and Google. Historically, Amazon has employed more women than Microsoft and Google also so they demonstrate a commitment to fair employment practices, which are crucial for evolution of technology ecosystems like cloud computing.
- What tutorial and project did you explore in GCP (or another platform), and what was your experience?
Azure is a top heavy mess just like Windows. GCP is certainly the most accessible and most emulates the Apple UX experience. AWS is the most responsible and cost structured.
Conclusion:
Even if these work well and have become the standard, they seem deficient and soulless. There needs to be a linux based cloud service like OpenStack, CloudStack, or NextCloud, and like LLMs there needs to be a common path toward building on a cloud platform that isn’t a monetized extension of flush companies that dominate the market. As taxpayers we are already subsidizing the networks, companies, and taxes of the executives, so I would like to see a universal basic cloud computing credit as a dividend on my investment as a taxpayer. Just like we should be getting internet, medical, water, and food discounted or comped because we already work and pay for the subsidies, grants, loans, tax breaks, and awards that built the cloud. Given this lopsided deal, it seems Amazon steals the least from the hostage consumer and gives you what you pay for without deceptive marketing.