Three deployment shapes
Different lenders need different operating envelopes. TransactIQ is delivered in three deployment tiers — self-hosted, managed India cloud, and dedicated private tenant — each designed around a specific fit profile. All three keep data inside the India data plane.
Self-hosted
Own the data planeTransactIQ runs inside the lender's own VPC on AWS, Azure, or GCP India regions. The customer owns the infrastructure, the data, and the operational envelope. Terra Insight delivers the software and the runbook; the lender operates it.
- ▸ Lenders with an existing SRE or platform engineering team
- ▸ Portfolios where data residency and control are binding constraints beyond basic compliance
- ▸ Volume profiles where unit economics of managed service are not the best shape
- ▸ Organisations with existing RBI-regulator exposure to in-house production systems
- ▸ Customer-owned Kubernetes / VM infrastructure inside their India region
- ▸ Versioned container images with documented upgrade cadence
- ▸ Runbook covering installation, upgrade, rollback, and operational troubleshooting
- ▸ Optional Terra Insight managed-upgrade assistance through early-access window
Managed India cloud
Fully managed on AWS MumbaiA multi-tenant managed service operated by Terra Insight on AWS Mumbai. The lender integrates via API; Terra Insight handles everything underneath. The fastest path from evaluation to production.
- ▸ NBFCs and digital lenders without a dedicated SRE team
- ▸ Teams optimising for go-live speed rather than infrastructure control
- ▸ Lenders where contractual SLAs from a vendor are preferable to in-house reliability burden
- ▸ Early-access partners validating production fit before negotiating enterprise terms
- ▸ Multi-tenant logical isolation with tenant-specific configuration
- ▸ Terra Insight owns uptime, patching, scaling, incident response
- ▸ SLAs documented in the Master Service Agreement
- ▸ Managed upgrade cadence coordinated with customer change windows
Private tenant
Dedicated single-tenant isolationA dedicated single-tenant instance on AWS Mumbai with isolated compute and storage. Terra Insight operates the deployment; the tenant retains dedicated resources. The right shape when multi-tenant isolation doesn't meet lender or regulator expectations.
- ▸ Enterprise NBFCs with strict single-tenant isolation requirements
- ▸ Regulated entities whose boards or auditors require dedicated infrastructure
- ▸ Lenders with volume or customisation profiles that exceed the managed-tier envelope
- ▸ Customers planning a self-hosted migration who want a managed starting point
- ▸ Dedicated data plane with no cross-tenant shared infrastructure
- ▸ Customer-specific networking posture and identity integration options
- ▸ Dedicated change windows, upgrade cadence, and incident-response paths
- ▸ Contractual portability clauses for migration to self-hosted if required
How to pick a tier
Four practical factors to decide which deployment shape fits the lender's operational reality. A shared conversation with Terra Insight engineering usually resolves this in one session.
Data-residency and control bar
If your board or regulator requires data to never leave the lender's VPC, self-hosted is the only shape. If managed AWS Mumbai is acceptable, managed or private tenant fit.
Engineering team shape
If you have an SRE team that already operates production systems, self-hosted is a good fit. If not, managed removes the reliability burden from a team that should not be carrying it.
Time-to-go-live
Managed is the fastest path — the integration window is days-to-weeks rather than months. Self-hosted adds infrastructure provisioning and operational readiness steps.
Escape velocity
All three tiers support documented data extraction and configuration export at contract end. Self-hosted has the shortest operational exit path because the customer already owns the data plane.
Ready to scope the right tier?
A 45-minute deployment-fit conversation with Terra Insight engineering typically narrows this to the right tier and surfaces any operational constraints that affect it.
Request deployment-fit call