Northbridge Systems cut its AWS bill from ₹6.4L to ₹3.9L a month
A 120-person software house was running client and internal workloads on an AWS estate that had grown project by project. We right-sized compute, cleaned up storage and network waste, and moved steady workloads onto Savings Plans.
- Custom software & SaaS development
- 120 people · Mid-size
- Bengaluru
- 5 weeks
Monthly spend
- Saved / month
- ₹2.5L
- Reduction
- 39%
- Payback
- 19 days
₹30L
Annual saving
129
Instances right-sized
9.4 TB
Orphaned storage removed
0
Client-facing incidents
The challenge
Northbridge had grown from 40 to 120 engineers in three years, and its AWS account had grown with it. Every new client project brought a fresh set of EC2 instances, RDS databases and S3 buckets, sized generously at launch and never revisited.
Leadership could see the monthly bill climbing past ₹6 lakh but had no clear view of which projects drove it. Tagging was inconsistent, staging environments ran around the clock, and nobody owned the shared networking costs.

What we reviewed
- 14 months of AWS Cost and Usage Reports
- 212 EC2 instances across 3 regions
- 38 RDS instances and their utilization
- S3 buckets, EBS volumes and snapshot history
- NAT gateway, load balancer and data-transfer patterns
- CloudWatch log groups and retention settings
The approach
- 1
Right-size compute to real utilization
Fourteen days of CPU and memory data showed 61% of instances averaging under 15% CPU. We mapped each to a smaller or newer-generation type (moving suitable workloads to Graviton) and staged the changes project by project to avoid any client impact.
- 2
Schedule non-production environments
Staging and QA environments were put on office-hours schedules, saving roughly 65% of their running hours with no change to developer workflow.
- 3
Clean up storage and snapshots
We removed 9.4 TB of orphaned EBS volumes and outdated snapshots, moved infrequently accessed S3 data to Intelligent-Tiering and set lifecycle rules on build artifacts.
- 4
Fix network waste
Consolidating NAT gateways, adding VPC endpoints for S3 and DynamoDB traffic and releasing idle Elastic IPs and unused load balancers cut network charges by 40%.
- 5
Commit only to the stable baseline
Once the estate was right-sized, we sized a one-year Compute Savings Plan to the new steady-state floor, so the discount applied to usage that would not shrink further.
Before and after
Monthly cost by line item, in Indian rupees.
Northbridge Systems — monthly cloud spend
Monthly| Line item | Before | After | Saving |
|---|---|---|---|
| EC2 compute (production + staging) | ₹2,85,000 | ₹1,62,000 | −₹1,23,000 |
| RDS databases | ₹1,18,000 | ₹74,000 | −₹44,000 |
| S3, EBS and snapshots | ₹72,000 | ₹41,000 | −₹31,000 |
| Data transfer and NAT gateways | ₹64,000 | ₹38,000 | −₹26,000 |
| Load balancers and idle Elastic IPs | ₹38,000 | ₹21,000 | −₹17,000 |
| CloudWatch logs and monitoring | ₹42,000 | ₹34,000 | −₹8,000 |
| Support plan and miscellaneous | ₹21,000 | ₹20,000 | −₹1,000 |
| Total | ₹6,40,000 | ₹3,90,000 | ₹2,50,000 (39%) |
Monthly spend over time
₹30L saved per yearTwo months before the engagement, then six months after the plan was delivered.
Timeline
Week 1
NDA, access and data pull
Read-only IAM role, Cost and Usage Reports and utilization exports.
Weeks 2–3
Analysis and review calls
Instance-by-instance recommendations validated with each project lead.
Week 4
Plan delivered
Prioritized plan with 46 actions, projected savings and rollback steps.
Week 5
Quick wins live
Schedules, storage clean-up and network fixes implemented by the client team with our support.
“We assumed our AWS bill was simply the cost of growth. beforetime showed us that almost 40% of it was idle capacity and old decisions. The plan was precise enough that our DevOps team could execute it in a sprint.”
Related case studies
All case studies
CloudMonthly spend
₹7.2L → ₹4.2L
Sentra Cloud · 110 people · Pune
Sentra Cloud saved 41% on Google Cloud with committed-use discounts and idle clean-up
Read case study
CloudMonthly spend
₹2.5L → ₹1.6L
Haven Systems · 24 people · Vadodara
Haven Systems saved 36% on AWS with Savings Plans and reserved databases
Read case study
CloudMonthly spend
₹8.4L → ₹5.7L
Orchid Softlabs · 105 people · Ahmedabad
Orchid Softlabs cut Azure and Microsoft 365 spend by 32% in one engagement
Read case studyFind out what your agency could save
Book a 20-minute call. We'll show you where the quickest wins usually are for a team like yours, with no data required and no obligation.
- Free 20-minute assessment
- NDA before any data is shared
- Fixed-scope quote, 30/70 payment terms
- 7 days of post-delivery support