Cost Optimisation

How to Reduce Your AWS Bill in India — 10 Proven Cost Optimisations

Indian businesses overspend on AWS by 30-60% due to over-provisioning, unused resources, and not using purchasing programs. Below are 10 high-impact AWS cost optimisations that Indian SMBs and enterprises can implement in 1-4 weeks. Combined savings typically exceed 40% of monthly AWS spend.

Optimisation 1 — Buy 1-year or 3-year Reserved Instances / Savings Plans: For workloads running 24/7 (your production database, main app servers), buy 1-year or 3-year Reserved Instances or Compute Savings Plans. Savings: 30-70% vs on-demand. Start with workloads you know are stable for 12+ months. For an Indian SMB spending ₹2 lakh/month on AWS, this alone can save ₹60,000-1 lakh/month.
Optimisation 2 — Right-size your EC2 instances: Most Indian businesses run instances 2-3x larger than needed. Use AWS Compute Optimizer (free in Console) to see which instances are over-sized. Downsize gradually — go from t3.large to t3.medium first, monitor for performance, then potentially to t3.small. Typical savings: 30-50% on compute.
Optimisation 3 — Use Spot Instances for fault-tolerant workloads: Spot Instances are 70-90% cheaper than on-demand but can be reclaimed by AWS with 2-minute notice. Perfect for: batch jobs, ML training, dev/test environments, stateless web servers behind load balancers. For Indian businesses running CI/CD pipelines or batch ETL, Spot can cut those costs to 10-30% of on-demand.
Optimisation 4 — S3 Intelligent-Tiering and lifecycle policies: Storage on S3 Standard is ~₹2/GB/month. S3 Glacier is ~₹0.32/GB/month (6x cheaper). Configure lifecycle policies: data accessed in 30 days stays in Standard; 30-90 days → S3 Standard-IA; >90 days → Glacier; >1 year → Glacier Deep Archive. For backup-heavy workloads, this saves 50-80% on storage.
Optimisation 5 — Delete unattached EBS volumes and snapshots: EBS volumes and snapshots persist even after EC2 instances are terminated, racking up unnecessary cost. Use AWS Trusted Advisor (Basic Support, free) or third-party tools to find: (a) EBS volumes not attached to any instance, (b) snapshots older than 90 days you don't need, (c) old AMIs. Deletion is free; savings are recurring.
Optimisation 6 — Reduce data transfer (egress) charges: AWS charges ~₹8/GB for outbound internet traffic from Mumbai region. Reduce by: (1) Using CloudFront CDN — cached responses count as CloudFront egress (~₹6/GB) vs origin egress, (2) Compressing responses (gzip, brotli), (3) Caching static content aggressively, (4) Moving customer-facing traffic to CloudFront. Data transfer optimisation can save 20-40% for B2C apps.
Optimisation 7 — Use Graviton (ARM) instances where possible: AWS Graviton (ARM-based) instances are 20-40% cheaper than equivalent x86 instances for the same performance. Most modern languages (Python, Java, Node.js, Go) run on Graviton with minimal code changes. For Indian businesses on managed services (RDS, Lambda), switch to Graviton variants — savings flow automatically.
Optimisation 8 — Stop dev/test environments off-hours: Indian businesses' dev/test environments often run 24/7 but are used only 9 AM - 7 PM × 5 days = 50 hours/week (~30% of the week). Use AWS Instance Scheduler to auto-stop instances after work hours and weekends. Savings: ~70% on dev/test compute. Negligible operational impact (instances restart at 9 AM Monday-Friday).
Optimisation 9 — Move to managed services from self-managed: Running your own MongoDB on EC2 typically costs ₹15,000-30,000/month (instance + storage + ops). AWS DocumentDB or DynamoDB equivalents cost ₹8,000-15,000/month at similar scale because you save operational overhead (replication, backups, upgrades). Counter-intuitively, managed services are often cheaper at total cost of ownership.
Optimisation 10 — Use AWS Budgets and Cost Anomaly Detection: In AWS Console > Cost Management, set up: (1) Monthly budget alerts at 50%, 75%, 100% of expected spend, (2) Anomaly Detection for unusual spikes (catches misconfigured Lambda, runaway crawlers, etc.). For Indian businesses, alerts save you from "₹10 lakh AWS bill surprise" months that happen with un-monitored cloud accounts.

Frequently Asked Questions

How much can an Indian business realistically save on AWS?

For businesses that have done minimal cost optimisation, 30-60% savings are typical in 4-8 weeks of effort. For more mature cloud users already using RIs and right-sizing, additional 10-20% savings via Spot, Graviton, and S3 lifecycle. Indian SMBs spending ₹5+ lakh/month on AWS should review costs quarterly — savings compound.

Should I buy Reserved Instances if my workload might change in 6 months?

Use Convertible Reserved Instances or Compute Savings Plans (more flexible than Standard RIs). Convertible RIs allow you to swap instance types (e.g., m5 to c5) without losing commitment. Savings Plans apply to any compute (EC2, Fargate, Lambda) within your hourly commit. These give 90% of RI savings with much more flexibility.

Is buying AWS through an Indian partner cheaper?

Not directly — AWS list prices are the same globally. But Indian AWS partners can: (1) issue GST invoice (vs USD direct from AWS for Indian businesses), (2) provide cost optimisation services, (3) negotiate enterprise discount programs (EDP) for spend above $500K/year. For most Indian SMBs, partner buying is mostly about GST invoicing, not direct price savings.

What's the fastest way to find AWS waste for an Indian business?

4-hour audit: (1) AWS Compute Optimizer — finds oversized instances (Day 1, 30 min). (2) AWS Trusted Advisor — finds unattached EBS, old snapshots, unused load balancers (Day 1, 30 min). (3) AWS Cost Explorer Reserved Instance recommendations (Day 1, 30 min). (4) S3 storage analytics — find rarely-accessed data (Day 1, 30 min). (5) Tag cost allocation by team/project (Days 2-3). Implementing the top 3 findings typically yields 20-30% savings in week 1.

Should Indian businesses consider moving from AWS to Azure or GCP for cost reduction?

Switching clouds is rarely cost-effective. The savings (typically 10-20% on equivalent compute) are usually consumed by migration cost (₹10-50 lakh) and operational disruption. The exception: workloads with specific GCP advantages (BigQuery for analytics, Vertex AI for ML) where the workload-specific savings justify migration. For most cost optimisation, doing it within AWS via RIs, right-sizing, Spot is far higher ROI than migrating clouds.

Free AWS cost audit for Indian businesses — identify 30-60% savings in 4 hours.