giip
SES Proposal
Oracle to AWS

How to migrate Oracle to AWS, and how to choose the target

The difficulty of an Oracle migration is set by where you are going. Staying on Oracle and moving to a different engine are two different projects with two different checklists.

This page is for teams that want to move on-premises Oracle to AWS but have not yet narrowed the target options. GIIP has migrated an on-premises Oracle environment with approximately 120,000 tables to AWS, and handles environments with extreme object counts.

Does this sound familiar?
!

Too many target options

Stay on Oracle, or move to another engine? You have nothing solid to decide with.

!

The volume of PL/SQL is unknown

Stored procedures and packages hold business logic, and it is unclear how much of it can move.

!

The inventory never finishes

With that many tables, views, synonyms and triggers, even producing the list of what must move is an unfinished project.

Why the decision is hard
01

Each target carries different premises

Staying on Oracle makes licensing and topology the issue; changing engine makes SQL and PL/SQL rewriting the issue. The same phrase "AWS migration" describes two different jobs.

02

Dependencies are buried in business logic

The longer an environment has run, the more business judgement sits inside triggers and packages. Each one is a decision: move it, or lift it out.

03

You cannot estimate what you have not counted

With a high object count, no mechanical extraction means no defensible estimate at all.

What GIIP does

Before laying out options, establish the scale and the dependencies mechanically.

Mechanical object inventory

Extract tables, views, synonyms, stored procedures and triggers per schema, producing both the scope list and the evidence that it is correct.

Frame the target options

From the volume of PL/SQL, licensing terms, external integrations and required availability, present the realistic options and their trade-offs.

Design the verification

Make before/after reconciliation of counts, definitions and dependencies something a machine runs across the whole scope.

Related pages

What you can check yourself first

  • Oracle version and edition
  • Number of schemas and approximate table count
  • Rough volume of stored procedures and packages
  • Database links and integrations with external systems
  • Number of batch processes and when they run
  • Terms of your current licence agreement
  • How many hours the business can be down

What we can evidence

SourceOn-premises Oracle environment
Table countApproximately 120,000 — not records, not databases
TargetAWS
Main challengeManaging a very large object count and its dependencies

The specific AWS service configuration used as the target is beyond what has been confirmed and is not published. Customer names and business data are likewise withheld.

Frequently asked questions

Should we keep Oracle on AWS?

The more business logic lives in PL/SQL, the lower the migration risk of staying on Oracle — but the licence cost stays too. The choice is as much a contractual decision as a technical one.

What makes a high table count a problem?

Not the migration itself, but proving the scope is complete. Manual inventory does not finish, so extraction and reconciliation have to be automated.

Is 120,000 a record count?

No. It is a table count — not records and not databases.

What will it cost?

It varies with scale and breadth of dependencies. We do not quote figures we have not established. The factors that drive cost are set out on a separate page.

Continue reading

Check whether your Oracle environment can be migrated

Tell us the version, number of schemas, approximate table count and external integrations.

contact@littleworld.net