India’s Leading Domain & Web Hosting Provider
Support
Domain & Login
Blog
Workload Placement: Choosing Cloud, Region or Edge for Each Workload
A workload placement assessment decides where each workload should run: AWS, Azure, Google Cloud, an edge or hybrid platform, colocation or its current data centre, and in which region. NTSPL scores every workload on latency, data residency, licence economics, service availability in Indian regions and operating cost, producing a placement matrix that CIOs can defend to the board and the auditor.
Delivered on
In scope
Why Default Placement Becomes Expensive
Most placement decisions are made by default: the cloud the first project used, or the one with the best discount offer. The result is Windows and SQL Server estates on a cloud where existing licences do not carry over, plant systems moved to a region 25 ms away from the line they control, and regulated data placed somewhere the compliance team cannot evidence. Fixing a wrong placement after go-live costs roughly as much as the original migration.
NTSPL builds a placement matrix with one row per workload and seven weighted criteria: latency tolerance, data residency and regulator constraints, licence portability, native service fit, Indian-region availability, skills, and three-year run cost. Hard constraints eliminate options first; weighted scoring ranks what remains. The answer may be one cloud for most workloads, but every exception is explicit and costed.
When Placement Needs a Formal Decision
Signals you need Workload Placement Assessment
- Two or more clouds are already in use and new workloads are placed by whichever team asks first
- More than 40 percent of the estate is Windows Server or SQL Server with Software Assurance
- Plant, store or campus systems need under 10 ms response to local equipment
- Customer contracts or a regulator specify data location, segregation or audit access by provider
When to choose something else
- The cloud provider is already contractually fixed with a committed spend: skip to readiness and wave planning on that provider.
- Fewer than 20 workloads with no regulated data: a one-day decision workshop is enough.
Placement Criteria We Score
Latency and locality
Measured round-trip times from sites to candidate regions and edge options, against each workload's tolerance.
Residency and regulation
DPDP Act 2023 obligations, RBI, SEBI or IRDAI constraints and client contract clauses on data location and segregation.
Licence economics
Windows Server, SQL Server, Oracle and RHEL entitlements, Hybrid Benefit eligibility and BYOL rules by provider.
Service fit and regional availability
Whether required managed services, GPU families or SAP-certified instances exist in Mumbai, Hyderabad, Pune, Chennai or Delhi.
Run cost and skills
Three-year cost per placement option, plus whether the team can operate it without adding headcount.
Constraint-then-Score Placement Model
Constraint-then-Score Workload Placement Model
Hard constraints remove impossible options for each workload, then weighted criteria rank the survivors. Exceptions to the primary platform are costed and approved individually.
Select any component
Constraint-then-Score Workload Placement Model: click any of its 7 components to see its role and the design choice NTSPL makes for it, or trace the 7-step primary path.
Hosting Options on AWS, Azure and Google Cloud
On AWS the placement question is often Mumbai versus Hyderabad versus a Local Zone or Outposts for latency-bound plant and store systems.
Azure scores strongly where Windows Server and SQL Server licences with Software Assurance carry over, and Azure Local covers workloads that must stay on site.
Google Cloud placement is weighed on Mumbai or Delhi, managed analytics and Kubernetes fit, and per-second billing for bursty workloads.
Placement Trade-offs With Thresholds
Single primary cloud or deliberate multi-cloud
Default to one primary cloud for at least 70 percent of workloads; two operating models double tooling, skills and security baselines. Add a second cloud only for a named reason: licence economics above roughly 20 percent saving, a service with no equivalent, or a client or regulator requirement. Record each reason in the exception log.
Region pairing in India
Choose primary and secondary regions inside India when residency matters: Mumbai and Hyderabad on AWS, Central India and South India on Azure, Mumbai and Delhi on Google Cloud. Check that every required service exists in both regions before committing; secondary regions sometimes lack newer instance families or managed services.
Edge versus cloud for site systems
Keep workloads at the site if they need under 10 ms to local equipment or must run through a WAN outage, such as MES, SCADA historians, parking or gate control. Move their reporting and integration tiers to cloud. Use hybrid platforms only when there are more than a handful of such servers per site.
Licence-driven placement
Where Windows Server and SQL Server cores with Software Assurance exceed a few hundred, model Azure Hybrid Benefit against BYOL on dedicated hosts elsewhere. Licence savings can outweigh compute price differences, but only if Software Assurance renewal is assumed for the full three-year model.
Four-Week Placement Study
Register and constraint capture
Workloads imported and hard constraints collected from compliance, legal and site operations.
Latency measurement
Probes from major sites to candidate regions and edge options for a full business week.
Scoring and licence modelling
Weighted scoring and licence portability tested per workload.
Costing and sign-off
Three-year cost for each surviving option; matrix and exceptions approved by owners and the CIO.
What you receive
How Placement Quality Is Judged
| KPI | Typical target |
|---|---|
| Workloads with a recorded placement rationale | 100 percent |
| Share on primary platform | 70 percent or more unless justified |
| Placements reversed within 12 months | Under 3 percent |
| Licence cost versus model after migration | Within 10 percent |
Placement Mistakes and Their Remedies
| Risk | How NTSPL prevents it |
|---|---|
| A discount offer decides placement | Credits are modelled separately and expire in year one of the three-year comparison. |
| A service missing in the secondary region | Every placement checks the service list in both regions before approval. |
| Multi-cloud sprawl through exceptions | Each exception needs a costed reason and is reviewed annually against the primary platform. |
Should we run on one cloud or split workloads across AWS, Azure and Google Cloud?
Run on one primary cloud unless a specific workload has a measurable reason to go elsewhere. A single platform keeps one landing zone, one security baseline, one skills track and one commitment discount. Splitting is justified when licence economics differ by more than about 20 percent, when a needed service has no equivalent, or when a client or regulator demands it. In our placement studies, the typical outcome for an Indian mid-market estate is 75-90 percent on one provider, a handful of exceptions on a second, and site-bound systems at the edge. Choosing multi-cloud as a principle rather than per workload usually raises run cost.
Key facts: Workload Placement Assessment
Workload Placement Assessment in Practice
2 Workload Placement Assessment scenarios, from Pharmaceutical contract development and manufacturing and Retail real estate (mall operations), plus related platform case studies.
Pharma CDMO splits 92 workloads across Azure, AWS and plant edge on contract rules
Mall operator keeps parking and CCTV on site and moves leasing to Azure South India
Audit firm moves 22 Windows servers to Azure and serves 180 staff via Virtual Desktop
Specialty chemicals group links 6 plants and HQ to AWS with dual-path hybrid network
Workload Placement Assessment: Frequently Asked Questions
How long does a workload placement assessment take?
Four weeks for up to about 300 workloads, provided discovery or application data already exists. Without it, add discovery first. One full business week is reserved for latency probes so placement uses measured, not published, round-trip times. Results remain valid until a major contract, regulatory or business change.
What does a placement assessment cost?
Usually 14-20 consultant days. The largest variables are the number of sites to probe and how many licence agreements need analysis. Costing three providers rather than two adds two to three days. Probe devices are shipped to sites pre-configured and returned at the end.
How does data residency affect placement in India?
Payment system data must be stored only in India under RBI's 2018 directive, and many RBI, SEBI and IRDAI-regulated workloads are expected to stay in Indian regions. Under the DPDP Act 2023, cross-border transfer is permitted except to countries the government restricts, but contracts may be stricter. We treat each as a hard constraint where it applies.
Can a workload stay on-premises in the final placement?
Yes, and often it should: latency-bound plant systems, hardware-locked applications or workloads close to retirement. Each retained workload gets a review date and, where useful, Arc or equivalent management so it is governed alongside cloud. Retained workloads stay in the matrix so nothing silently drops out of scope.
Do provider credits and discounts change the answer?
We model them separately. Credits usually cover year one only, so a three-year comparison shows whether they change the answer or just the first invoice. Credits that require long commitments are also tested against the full three-year horizon.
Can the placement rules be reused for new projects?
Yes. We publish the constraints, weights and thresholds as a short standard, so new workloads are placed in a 30-minute review rather than a new study. The review checks hard constraints first and only scores the exceptions.