Back to UKOUG Discover: EBS Unleashed and Oracle Database Services Decoded

Next week I’m heading back to UKOUG Discover 2026 Conference, my fourth time presenting at UKOUG.

This year Discover moves from Birmingham to London, November 23–25, 2026, at Novotel London West in Hammersmith, with Super Monday on the 23rd and the full conference on the 24th and 25th.

Honestly, the part I look forward to most is seeing my UKOUG friends again. This community has been generous to me. Earlier this year, UKOUG recognized me with a Speaker of the Year award, and I’m truly humbled by it. Thank you to everyone who came to my sessions, challenged my assumptions and asked the hard questions, and to UKOUG for the recognition.

That’s what makes coming back so easy.

I have two sessions this year, one on Super Monday and one on Tuesday afternoon. They cover very different subjects, but they are built around the same principle:

Evidence first. Product names last.

That theme comes from a pattern I keep seeing in customer conversations. Most organisations don’t need another catalogue of Oracle cloud services. They need help making two much harder decisions:

What is the right modernization path for an EBS estate that may be around for another decade?

And:

Which Oracle Database service actually fits each workload?

Those are the two questions I’m taking to Discover this year.

Oracle EBS Unleashed: Unlocking Performance, Savings, and Resilience on OCI and Multicloud

Monday 23 November · 11:00 · Mezzanine – Pouilly

Let me start with something that sometimes gets lost in all the discussion about application modernization:

EBS is staying.

Oracle E-Business Suite 12.2 has Premier Support through at least 2037. For many organisations, therefore, the immediate question isn’t whether to replace EBS. The more practical question is:

Where should EBS run for the next five or ten years, and how do you get it there safely?

That sounds like an infrastructure question. It isn’t.

A successful move can improve performance, eliminate aging hardware, strengthen HA and DR, and change the economics of running EBS. It can also give organisations choices across OCI and Oracle Database services in Azure, AWS and Google Cloud.

But none of those outcomes are automatic.

They are switched on or off by architecture, sizing, licensing and migration decisions made before the move begins.

In the session I’ll focus on four traps I repeatedly see:

  • Moving without a horizon. Designing for two years and designing for ten years should not produce the same architecture or investment.
  • Sizing from installed hardware instead of workload evidence. If your current servers are oversized, moving that footprint to the cloud simply preserves the waste and may preserve an unnecessarily large licence position with it.
  • Separating the application and database tiers without understanding latency. EBS is sensitive to the conversation between those tiers. Cloud placement matters.
  • Treating cutover and DR as activities instead of tests. Your production cutover should not be the first time you prove that the migration works.

One anonymised EBS workload, including a two-node RAC database sized from its actual AWR history, runs through the session.

Instead of beginning with “Which Oracle product should I buy?”, we’ll begin with the workload.

We’ll look at when lift-and-shift makes sense, when re-platforming onto Base Database, Exascale or Exadata changes the equation, and how multicloud fits when your application landscape already has gravity in Azure, AWS or Google Cloud.

We’ll also look honestly at what Exadata fixes and what it doesn’t.

Exadata can dramatically change the economics of I/O-intensive database workloads. It is not a cure for bad SQL. Smart Flash Cache may help a workload performing billions of physical reads. It cannot rescue a DELETE statement with no WHERE clause.

We’ll connect that architecture discussion to resilience, licensing and cutover: what HA actually protects, how UK region availability influences DR design, how BYOL changes the economics, and why I want multiple rehearsals before anyone calls a migration plan “ready.”

The objective isn’t to leave you with another cloud product comparison.

You’ll leave with an eight-step EBS modernization roadmap, with an owner against every step. The first few steps are deliberately inexpensive because the best time to discover that your sizing, licensing or migration assumptions are wrong is before you’ve provisioned anything.

Last year in Birmingham I talked about resilient EBS on OCI. This year we’re taking the next step and looking at the whole journey.

Oracle Database Services Decoded: OCI and Multicloud

Tuesday 24 November · 15:00–15:45 · Mezzanine – Beaujolais

Here’s another statement that may sound slightly uncomfortable at an Oracle conference:

Autonomous is not always the right answer. Neither is Exadata.

Oracle customers now have more database deployment choices than ever across OCI and multicloud services.

More choice is good.

But it has made one apparently simple question surprisingly difficult:

Which database service is actually right for this workload?

My Tuesday session is about answering that question without starting with a product name.

The underlying Oracle Database technology may be familiar, but the service around it radically changes the operating model: who patches, how you scale, how much control you retain, what infrastructure you manage, what you pay and which licensing options are available.

So instead of starting with Base Database, Exascale, Exadata or Autonomous, we’ll start with five questions:

  1. Performance shape — Is the workload latency-sensitive, throughput-heavy, bursty or steady? What actually hurts today?
  2. Control need — Do you need to control patch timing, database versions, configuration, operating-system agents or network placement?
  3. Licence position — What Oracle licences do you already own, which editions are they, and where can you actually use them?
  4. Operating capacity — How much infrastructure and database operations work should your team still be doing?
  5. Cloud gravity — Where do your applications, data, users and existing cloud commitments already live?

Only after answering those questions should the product names appear.

I’ll show what this looks like with real evidence.

One example is an anonymised Oracle Cloud Premigration Advisor Tool assessment of a 19c Standard Edition 2 database against Autonomous Database Serverless. The assessment ran 70 checks. Twelve came back as Action Required.

That’s exactly why these assessments matter.

Let the evidence eliminate the wrong platform before your users have to.

We’ll also look at the economics. I’ll use Oracle’s own pricing information to unpack the differences between Base Database, Exascale, Exadata Dedicated, Cloud@Customer and Autonomous, including some licensing scenarios where the obvious answer isn’t necessarily the cheapest answer.

And we’ll talk about something else I think matters as Oracle database services become increasingly automated:

Automation removes toil, not accountability.

A DBA may stop patching infrastructure or performing some routine platform tasks. But ownership of SQL, schemas, performance, data security, architecture and cost doesn’t disappear.

The job changes. The responsibility doesn’t.

We’ll finish with the multicloud question. Database@Azure, Database@AWS and Database@Google Cloud create some very interesting architecture opportunities, but service and regional coverage matter, particularly when you’re trying to build a UK HA or DR strategy.

The goal isn’t to declare one database service the winner.

It’s to give you a framework that helps you explain why this workload belongs on this service.

You’ll leave with the framework on one page and three things to do when you return to work: one for the DBA, one for the architect and one for the IT leader.

Come Find Me

If you’re wrestling with an EBS modernization decision, trying to choose between Base Database, Exascale, Exadata or Autonomous, or working out where Database@Azure, Database@AWS or Database@Google Cloud fits, bring the difficult case.

I don’t promise that the answer will always be a particular Oracle service.

That’s rather the point.

And if we’ve met at Discover before, please say hello. The best conversations usually start after the slides are finished.

Monday 23 November, 11:00, Mezzanine – Pouilly

Oracle EBS Unleashed

Tuesday 24 November, 15:00–15:45, Mezzanine – Beaujolais

Oracle Database Services Decoded

Connect: LinkedIn

Fourth time at Discover. New city, same friends.

See you in Hammersmith.