AWS Cloud Migration in Delhi

What Delhi businesses need from this

The most researched market here. Buyers arrive with a requirements document and a shortlist of three to five vendors. Terms like migration and compliance carry weight, and a written answer beats a lower quote.

What is included

• Cost modelling before the move, covering data transfer, storage class selection, NAT gateway processing, public IPv4 addresses and log volume
• Region selection for Indian users and Indian data residency, with the latency and compliance trade-off stated explicitly
• Cutover with a rollback position, often a short parallel run rather than a single high-stakes change window
• Post migration tuning: right sizing, scheduled scaling, budgets, and removal of the resources the assessment flagged

Approaches and how they differ

• Rehost (lift and shift): Move the workload as it is onto EC2 with minimal change. Fastest route to AWS and a good first step for leaving old hardware, but the technical debt and much of the running cost travel with it.
• Replatform: Move onto a supported managed runtime such as RDS for the database and S3 for file storage. Cuts server administration without rewriting the application, and is the usual middle position for an Indian SME.
• Refactor: Rebuild around managed services and serverless components. Removes server administration entirely but creates genuine AWS dependency, and is only justified when the application will be actively maintained.
• Retire and replace: The legacy application is not migrated at all. Correct when source code is unavailable, the platform is unsupported, or the maintenance cost of the old system exceeds the cost of a replacement.

What it suits, and what it does not

• Good fit: on-premises servers, VPS instances or a colocation contract reaching end of life, with a known inventory
• Good fit: applications with a stated recovery objective, where a tested restore in a second availability zone is required
• Good fit: bursty workloads that idle cheaply under EC2 Auto Scaling, instead of being sized for a peak that occurs twice a year
• Poor fit: a single simple website or mail server, where managed hosting costs a fraction of an AWS account and its bill shape
• Poor fit: flat, steady compute, where a managed platform is normally cheaper than running instances continuously

Where projects like this go wrong

• Migrating without an inventory, so dependencies on on-premises databases, file shares and internal DNS are discovered during cutover
• Underestimating charges outside the instance rate: cross availability zone traffic, internet data transfer, and the address and logging overhead billed alongside compute
• Choosing a region for a promotional offer rather than for latency to Indian users and Indian data residency requirements
• Adopting managed services early for convenience, then finding the application cannot move to another provider without rework
• Leaving orphaned resources after the old environment is decommissioned, which is the most common reason a bill does not fall after a migration

Who is responsible for what

• The client owns the AWS account, root credentials, MFA enrolment and billing. The provider works through a defined IAM role and never as account owner
• The client owns IAM policy decisions. The provider can draft roles, but approval of who can do what remains with the client
• The provider handles the migration build, cutover and ongoing technical operation of the platform
• The client classifies the data and states where it may be stored, including whether it must remain in India
• The client approves the cost model and budget thresholds. The provider configures them, but the spending decision belongs to the client

What buyers ask before starting

• How long will the migration take, and can the old system keep running throughout?
• What will this cost in AWS once we have moved?
• Which region should an Indian customer use?
• If it goes wrong can we move back?
• Do we have to rewrite the application to move?

What actually matters

• Total monthly cost including transfer, storage class, logging, backups and idle resources, compared honestly against the current bill
• How changeable the application is, which decides rehost against refactor before any AWS work begins
• Rollback. A move with a working way back is a project; a move without one is a bet
• Who owns the account and credentials afterwards, and whether the client's own team can operate the environment
• Data residency and contractual terms, where the customer is in India and some services are only offered in particular regions

Delhi specifically

The widest business spread of the four and the most competitive vendor market. South and central districts carry offices, trade houses and markets, outer colonies carry retail and clinics, older streets carry family firms.

Who buys this here

• Retail chains and franchise networks
• Fintech and SaaS firms with compliance rules
• Media firms with real traffic
• Trading houses moving off legacy platforms
• Enterprises consolidating servers

What they need it to do

• AWS and Azure migration scoped in phases
• Data residency, encryption, audit logs and access control
• Multi-zone design justified against its cost
• One contract covering compute, backup and monitoring
• Cost tagging that makes a departmental bill explainable

Connectivity and network conditions

Demand is larger and burstier: launches, sale season and always-on consumer platforms produce load no shared plan survives. Indian regions keep domestic latency low and rise sharply for overseas visitors.

How they tend to buy

Price is discussed openly and proposals compared line by line, but risk, response time and a written scope decide it. Buyers expect GST invoicing, a security questionnaire answered before kickoff and a named contact.

Areas covered

• Nehru Place
• Okhla Phase 1, 2 and 3
• Lajpat Nagar and Central Delhi
• Karol Bagh and Punjabi Bagh
• Rohini
• Dwarka
• Janakpuri and Paschim Vihar

Local detail that affects this work

• One visit across the spread can consume a working day
• Nehru Place and Okhla concentrate IT supply and spares
• Vendor density is high, so buyers arrive shortlisted
• Largest pool of enterprise migration work here
• Central Delhi buyers bring procurement rules early

Servers to rent rather than services to plan

If the requirement is a server rather than migration or advice, the three published plans are 2 vCPU, 4 GB, 25 GB SSD at ₹999 per month, 4 vCPU, 8 GB, 100 GB SSD at ₹1,799 and 6 vCPU, 16 GB, 150 GB SSD at ₹3,499.