Sovereign Cloud vs Global Hyperscalers: Data Residency Checklist for Indian SMBs
Hyperscaler or sovereign cloud? A practical checklist for Indian SMBs to map data, cut residency costs, and meet DPDP-era localisation demands.

Last year a 22-person textile exporter in Tiruppur told me their auditor had flagged a line item they didn't expect: a vendor onboarding form asked, in plain language, where their customer payment records were physically stored. They had been running their ERP on a hyperscaler region in Singapore for three years. Nobody had asked them that question before, and their answer, "Singapore, I think, and maybe a backup in Mumbai," was not good enough for a buyer who had recently tightened its supplier data policy. The fix took five weeks and cost more than it should have, because nobody had mapped their data before choosing a cloud.
That story is becoming routine. With the Digital Personal Data Protection Act, 2023 and its rules rolling out, sector regulators tightening localisation expectations, and a growing number of enterprise and government buyers writing residency clauses into their procurement, Indian SMBs are being asked where their data lives in a way they rarely were five years ago. At the same time, sovereign cloud offerings from Indian providers and in-country regions from the big hyperscalers have multiplied, and the pricing differences between them are wide and often misunderstood.
This post is a practical guide to what data residency actually requires, how to classify your data before you spend a rupee on infrastructure, how to compare hyperscaler regions against sovereign options on total cost, and where small businesses routinely overpay. Along the way we'll walk through a worked migration example and a checklist you can take to your vendor. If you're searching for a sovereign cloud India SMB data residency plan that doesn't blow up your budget, this is the sequence I'd follow.
Key Takeaways
- Data residency is a classification exercise first and an infrastructure decision second. Map your data categories before comparing providers.
- "In India" and "sovereign" are not the same thing. Check where backups, logs, support access, and replicas live, not just the primary database.
- Hyperscaler in-country regions often give you residency without full sovereignty. Sovereign providers add operational and legal controls at a premium that is sometimes worth paying and sometimes not.
- Most SMBs only need a small slice of data localised. Splitting workloads can cut the residency bill by half or more.
- Egress charges and cross-region replication quietly erase savings. Model them before signing.
- Get residency commitments in writing, with named regions and subprocessors, not marketing pages.
What does data residency actually require for an Indian SMB?
Start with the word itself. Data residency means the physical location where data is stored at rest. Data localisation is a broader legal requirement that certain data must be stored or processed inside a jurisdiction. Data sovereignty goes further and asks which legal authority can compel access to the data and under what conditions. These three ideas get blended constantly in vendor decks, and that blending is where SMBs get confused.
For most Indian SMBs the practical obligations come from a handful of sources:
- DPDP Act, 2023 and rules: Governs personal data of Indian data principals. It allows cross-border transfer by default except to countries the government restricts, but significant data fiduciaries and sector regulators can impose localisation. Most SMBs are not significant data fiduciaries, but your customers may be, and their contracts will flow obligations down to you.
- RBI payment data guidelines (2018 directive): Payment system data must be stored only in India. If you're a fintech, a payment aggregator's technology partner, or handle payment data in any form, this is non-negotiable. Foreign processing is allowed only with data returned to India within a stated window.
- IRDAI and SEBI requirements: Insurance and capital market intermediaries face localisation expectations for specific records. Check with your regulator's circulars, not a blog post.
- Government procurement and customer contracts: Increasingly, government departments, PSUs, and large enterprise buyers write residency into supplier agreements. This is often the most binding obligation an SMB encounters, because it shows up in a purchase order, not a statute.
- GST and corporate records: There is no blanket rule that GST invoices must be hosted in India, but you must be able to produce records on demand under the GST and Companies Act regimes. Where those records sit matters less than whether you can retrieve them quickly and in usable form.
The distinction that matters: residency asks "where is it stored," sovereignty asks "whose laws can reach it," and compliance asks "can you produce it on demand." A provider can satisfy the first and fail the second or third.
How should I classify my data before choosing a cloud?
This is the step most SMBs skip, and it's the reason they overpay. Spend a day, not a month, building a simple classification register. I use four tiers, which fit almost every SMB I've worked with:
- Regulated or contractually restricted: Payment records, customer personal data covered by a client contract, government project data, health or insurance records. These are the candidates for in-country, sovereign-grade hosting.
- Personal and confidential but unregulated: Employee HR records, internal financials, customer support tickets with personal details. Usually fine in an Indian region of a hyperscaler with good access controls.
- Internal operational: Project documents, email, collaboration files. Residency rarely matters here beyond what your tenancy agreement with Google Workspace or Microsoft 365 already offers.
- Public or derived: Website content, marketing assets, anonymised analytics. Host wherever is cheapest and fastest.
For each category, write down the system that holds it, the storage location today, the backup location, and who has administrative access, including vendor support staff. That last column catches more problems than all the others combined.
Pro Tip: Check where your backups and logs actually live
The most common residency breach I find is not the primary database. It's the backup vault in a different region that someone enabled for disaster recovery in 2021, plus application logs shipped to a SaaS monitoring tool headquartered abroad that stores PII in request payloads. Ask your engineer to list every service that touches regulated data, including observability and error tracking. You will almost certainly find at least one surprise.
Hyperscaler India regions versus sovereign providers: which should an SMB pick?
Here is where the real decisions are made. Indian options broadly fall into three groups: hyperscaler regions (AWS Mumbai and Hyderabad, Azure Central and South India, Google Cloud Mumbai and Delhi), Indian-owned sovereign or government-aligned cloud offerings, and managed private cloud providers that run on in-country data centres. The right choice depends on which of your categories are actually regulated and how much operational control you need.
| Criterion | Hyperscaler India region | Indian sovereign cloud offering | Managed private cloud (in-country DC) |
|---|---|---|---|
| Data stored in India | Yes, for selected services and regions | Yes, by design | Yes, by design |
| Service breadth | Very high, 200+ services | Narrower, core compute, storage, DBs | Depends on provider, usually VM-based |
| Support access from abroad | Possible; check contractual terms | Typically restricted to India-based staff | Usually restricted |
| Indicative cost for 4 vCPU, 16 GB VM with storage (monthly, INR) | ₹9,000 to ₹14,000 on-demand, lower with commitments | ₹11,000 to ₹18,000, varies widely | ₹12,000 to ₹20,000 including management |
| Best fit for | SMBs needing breadth and talent availability | Government-linked or strictly regulated SMBs | SMBs with legacy apps and tight sovereignty needs |
Treat those cost ranges as orientation, not quotes. Prices change, discounts vary sharply by commitment, and you must price your own workload. The point of the table is the shape of the trade-off: sovereign and managed private options usually cost more per unit of compute and offer narrower services, while hyperscaler regions are cheaper per unit but require you to verify support access and metadata handling in the contract.
A worked example: Gurgaon logistics firm
Consider a 15-person freight forwarding company in Gurgaon. They ran a booking system, a customer portal, a document store for bills of lading, and a small reporting database on three on-premise servers in a rented rack. Their hardware, power, bandwidth, and a part-time sysadmin cost them roughly ₹45,000 a month. Two of their larger customers now required that shipment and customer data remain in India, and one required that support access be limited to Indian personnel.
We classified their data. Shipment documents with consignee personal details and billing records fell into tier one because of a customer contract. The portal and booking system were tier two. Email and internal files were tier three. Marketing content was tier four.
Instead of moving everything to a sovereign provider, we split the estate. The tier-one document store and billing database went to a managed Indian private cloud environment with India-only support access, at roughly ₹22,000 a month. The booking system and portal went to the Mumbai region of a hyperscaler on reserved instances, around ₹8,000 a month after commitment discounts. Email and collaboration stayed on Google Workspace with its India data residency option, roughly ₹3,500 a month for the team. Backups were kept inside India using a region-pinned vault, adding about ₹2,000. The total came to roughly ₹35,500 a month, compared with ₹45,000 before.
The savings were not dramatic in the first month, because the migration itself cost them about ₹1.8 lakh in engineering and parallel running. But they retired the rack, eliminated a hardware refresh they had budgeted for at ₹6 lakh, and ended the contract dispute. Once you account for avoided hardware, the payback period was under nine months. The number that mattered most, though, was that their largest customer signed the renewal without the residency addendum they had been threatening.
If you're after a complete picture of cloud migration effort and managed support options, our cloud migration and managed services team works through exactly this kind of split, and our IT consulting engagements usually begin with the classification exercise described above.
Where do Indian SMBs overpay for data residency?
Three patterns come up repeatedly.
They localise everything. Putting marketing assets and internal wikis in the most expensive sovereign environment is pure waste. Localise tier one, consider tier two, and leave the rest where it's cheapest.
They ignore egress. Moving data out of a cloud costs money, and so does moving it between regions. A reporting system that pulls a few terabytes monthly from a Mumbai database into a Hyderabad analytics service can generate a surprising bill. Our post on cloud egress fees in India walks through how to model and reduce this, and it's worth reading before you finalise architecture.
They buy sovereignty they don't need. A sovereign private environment with India-only operators makes sense if a contract or regulator requires it. If your only driver is a vague sense that "sovereign is safer," a well-configured hyperscaler India region with customer-managed encryption keys held in India often meets the requirement at lower cost. Our companion piece, sovereign cloud vs public cloud for Indian SMBs, goes deeper on the cost curve.
Common Mistake: Assuming the region badge settles residency
Selecting "Mumbai" in a console does not guarantee that every service, managed control plane, support ticket attachment, or telemetry export stays in Mumbai. Some services are global by design. Identity, certain CDN features, and some AI and analytics services process data outside the region or replicate metadata globally. Ask the vendor for a service-by-service residency statement for the exact services you use, and retain it in your compliance file.
What is a practical data residency migration process?
Whether you do this yourself or brief a partner, this sequence works:
- Inventory systems and data flows. List every application, database, file store, backup, log pipeline, SaaS tool, and integration. Note where each lives today.
- Classify data into tiers. Use the four-tier model above. Get sign-off from whoever owns the customer contracts, not just IT.
- Pull the contractual and regulatory obligations. For each tier-one category, record the clause, the regulator, and the specific requirement (storage, processing, support access, return timeline).
- Shortlist providers by requirement, not by brand. Eliminate options that cannot meet tier-one requirements before comparing prices.
- Price the split architecture. Model compute, storage, backup, egress, support, and migration labour for at least two scenarios, including one with a full sovereign environment.
- Run a pilot on a non-critical workload. Test the residency controls, backup vault location, key management, and support access workflow before moving regulated data.
- Migrate tier one first, then tier two. Keep parallel running for at least two billing cycles so you can validate costs and performance.
- Document and audit. Store the residency statements, architecture diagrams, and access logs in a location you can produce on demand. Review quarterly.
Timelines vary, but for a 15 to 30 person SMB with a modest estate, expect six to ten weeks from classification to tier-one cutover if you start with clean documentation. Add three to four weeks if your legacy systems have undocumented integrations, which most do.
What checklist should I take to my vendor?
Bring these questions in writing and require written answers:
- Which specific services and regions will host our tier-one data, and can you provide a service-level residency statement?
- Where are backups, snapshots, replicas, logs, and support attachments stored?
- Who has administrative or support access, from which countries, and under what approval process?
- Are encryption keys held in India, and can we hold customer-managed keys?
- What is the contractual process if a government or court requests access, and will you notify us?
- What are egress and inter-region transfer charges for our projected volume?
- What is the exit plan, including data return format and timeline?
- Will you sign a data processing addendum that names Indian residency and subprocessors?
Vendors who answer these fluently are usually worth shortlisting. Vendors who redirect you to a marketing page usually are not.
Frequently asked questions
Do Indian SMBs legally need to store all their data in India?
Not generally. Most SMBs have no blanket legal localisation requirement. Obligations come from sector regulators such as the RBI for payment data, from specific contracts with customers or government buyers, and from the DPDP Act where the government restricts transfers. Classify your data to find out which obligations actually apply to you.
Is a hyperscaler's Mumbai region enough for data residency?
For many SMBs, yes, provided the services you use keep data in the region and the contract covers support access and metadata. It is not automatically sovereign, meaning foreign legal authorities may still be able to compel access through the vendor. If your contracts require sovereignty, you need a provider that commits to that in writing.
How much more does a sovereign cloud cost than a public cloud in India?
For comparable compute, sovereign and managed private options often run 20 to 60 percent higher, though the gap varies by commitment level, service breadth, and support model. The right comparison includes avoided hardware, staffing, and compliance risk, not just the monthly invoice.
Do I need to move my email and collaboration tools for residency?
Usually not as a first step. Google Workspace and Microsoft 365 offer India data residency options for many customers. Check your tenancy settings, confirm backup and retention locations, and verify that your residency commitment covers the specific data types you hold.
Can I split my data between an Indian sovereign provider and a global hyperscaler?
Yes, and for most SMBs a split architecture is the most cost-effective approach. The complexity lies in integration, latency, and egress, so model those costs and keep tier-one data within a single residency boundary as far as possible.
How long does a data residency migration take for a small business?
For a 15 to 30 person company with documented systems, six to ten weeks to cut over tier-one data is realistic. Undocumented integrations and legacy on-premise dependencies extend that timeline, sometimes significantly.
What happens if I don't address data residency and a client asks later?
You may lose renewals, face contractual penalties, or be unable to bid on certain tenders. Retrofitting residency under deadline pressure is almost always more expensive than planning it, as the Tiruppur exporter learned.
Conclusion: plan residency as a classification decision
The sovereign cloud India SMB data residency question is less about choosing the right logo and more about knowing your own data well enough to avoid overbuying. Classify first, split where you can, price egress and support access honestly, and demand written residency statements from anyone you hire. Done in that order, most small businesses can meet their obligations for less than they expected.
If you'd like a second pair of eyes on your data map or a costed split-architecture proposal, eDarpan can help with the IT consulting assessment and with the implementation through our cloud and managed services practice. You can also reach our team to talk through your specific obligations, and learn more about how we work on the about page. For broader context on the market shift, read what to check before switching to sovereign cloud and where Indian SMBs should start their cloud shift.
Image credit: Innovate Maryland Emerging Technology Center by MDGovpics via flickr (BY 2.0), sourced through Openverse.
Written by
Amit Verma
Cloud architect specializing in AWS, Azure, and GCP infrastructure. Amit has designed multi-region deployments for Indian enterprises and writes about cloud migration, cost optimization, and DevOps best practices.
Looking for a technology partner?
From IT consulting to virtual office to custom software — eDarpan can help.
Continue reading
Sovereign Cloud vs Public Cloud for Indian SMBs: Real Cost Trade-offs
Sovereign or public cloud? Real cost trade-offs for Indian SMBs, from hidden egress fees to DPDP compliance, with a practical audit to decide before you migrate.
Sovereign Cloud Is Live in India: What SMBs Should Check Before Switching
A managed Indian sovereign cloud is live. Before your SMB switches, learn the difference between residency, sovereignty and compliance, and the hidden costs to check.

Cloud Egress Fees in India: Cut the Data Transfer Bill You're Overpaying
Your cloud bill may be hiding a data transfer charge. Learn how cloud egress fees in India work, where they hide, and how to cut them this week.