Featuring Jan Kinkle (Dbvisit) and Edward Mace (Xynomix)
For the First Time, Sub-Second RPO Is Available on Oracle Standard Edition
For Oracle Standard Edition users, the disaster recovery landscape has changed
Want to see how this works in practice? Dbvisit and Xynomix are hosting a technical webinar, Zero Data Loss in Oracle: From Architecture to Real-World Recovery, where we’ll break down the Oracle redo flow, compare the different approaches to reducing data-loss exposure, and show StandbyMP Zero Data Loss in action through a real proof-of-concept use case.
With the launch of StandbyMP Zero Data Loss (ZDL), Oracle SE environments can now achieve sub-second RPO using continuous redo replication alongside established archive-log replication. It brings a level of data-loss protection that has traditionally pushed organisations towards Oracle Enterprise Edition and Data Guard, or more complex custom approaches.
For years, that was one of the accepted compromises of Standard Edition.
Oracle SE can be a very capable platform for organisations that do not need the additional functionality of Enterprise Edition. But disaster recovery has always complicated that decision. Without a native equivalent to Data Guard, organisations have typically relied on backups, custom scripting, archive-log replication or third-party solutions.
Those approaches can provide effective DR. The challenge has been getting the Recovery Point Objective (RPO) down from minutes to seconds.
As Jan Klinke, DR Solution Consultant at Dbvisit, recalls:
“The general notion is that if you are running Oracle Database Standard Edition you have to accept certain trade-offs. And the data loss was one of those trade-offs.”
For some workloads, an RPO measured in minutes is still perfectly acceptable.
But the practical reality is that data worth keeping is usually data worth protecting. And as increasingly important workloads have remained on Oracle SE, the question has shifted from whether those databases need DR to something more useful:
How much committed data can the business realistically afford to lose?
When five minutes actually matters
Ed Mace, Sales Team Leader at Xynomix, sees that question from the customer side. Xynomix provides Oracle managed services, consultancy, business continuity and disaster recovery support, so Ed is regularly involved in discussions around RPO, RTO and what recovery actually means to the business.
As Ed puts it:
“When I speak to customers, what matters is what that data represents to their business, how much of it they could lose, and whether they could realistically recreate it. Five minutes of lost transactions might be manageable in one environment, but in another it could mean lost payments, orders or operational activity with a very real financial impact.”
That is an important distinction.
A standby can be healthy and functioning exactly as designed while still being several minutes behind the latest committed transaction.
As Ed sees it, customers are becoming much more aware of that difference:
“The world’s changed in terms of the criticality of these systems. A lot of customers are now driven around their recovery time and recovery point objectives.”
And historically, if that recovery-point requirement became demanding enough, the answer was often Enterprise Edition.
That can make perfect sense where a workload genuinely needs Enterprise Edition. But Ed also works with organisations that deliberately use Standard Edition because they simply do not need those additional capabilities.
The question was whether those organisations could get a significantly lower RPO without changing edition purely for disaster recovery.
Closing the gap before the archive log
That was the technical problem Dbvisit set out to solve.
Traditional archive-log replication waits for redo to become an archive log before it is transferred and applied to the standby. That model is proven and reliable, but it leaves a gap between a transaction being committed and that transaction becoming recoverable elsewhere.
To move from minutes into fractions of a second, Dbvisit had to start protecting the redo earlier.
Jan explains:
“It was obvious that we couldn’t do it with the help of archive logs. It became clear that we had to somehow synchronize the online transaction log to the standby.”
ZDL monitors the active online redo at the filesystem level, captures changes as they are written and continuously synchronises them to the standby.
The engineering challenge was not simply moving that data quickly. The replicated redo also had to remain consistent and usable throughout the process.
The important change is simple:
You no longer have to wait for redo to become an archive log before you start protecting it.
Why not replace archive-log replication?
Because Dbvisit deliberately wanted to preserve the proven replication architecture underneath it.
Archive-log replication continues as before. ZDL adds a second, independent replication stream alongside it.
That gives the environment two complementary protection mechanisms. If network conditions change or a workload suddenly produces more redo than the continuous stream can accommodate, the established archive-log replication remains in place.
Jan describes the thinking simply: “When you work with Oracle databases you learn to be conservative.”
Ed sees that as an important part of the proposition: “Where do you go with an already good product? The standby capabilities were phenomenally reliable and good. This just enhances it even further.”
However, here is an important technical distinction here. StandbyMP Zero Data Loss should not be confused with synchronous Data Guard Maximum Protection.
Jan is very clear: “The truth is that it’s not exactly zero. The exact zero data loss can only be achieved with Data Guard in the Maximum Protection mode.”
For an unplanned failure, ZDL is designed to deliver sub-second RPO in suitable environments, with the architecture discussed by Jan operating at approximately a quarter of a second to one second.
So, do you actually need it?
Not every Oracle SE database needs sub-second RPO. A more useful question than “Can we achieve sub-second RPO?” is: “What would losing the last five minutes of this workload actually mean?”
For one system, the answer may be very little, but another, it could mean missing payments, orders, financial transactions or operational data that teams then have to reconstruct before the recovered system can be trusted.
Ed sees this as a cost-versus-risk discussion. The DBA can implement the technology, but the acceptable RPO is ultimately determined by what the organisation is prepared to lose.
And that is why this conversation goes beyond software.
From replication to recovery readiness
Xynomix’s role does not stop at implementing the technology.
Its Oracle DBAs can design and configure the environment, monitor production and standby systems, carry out test switchovers and provide ongoing support through its Disaster Recovery as a Service (DRaaS) offering.
In our upcoming webinar, Dec 1st, Dbvisit and Xynomix will take this discussion beyond the theory.
We’ll follow a confirmed transaction through the Oracle redo flow, compare Data Guard, ZDLRA, Oracle Cloud recovery options and StandbyMP Zero Data Loss, and look at the technical, operational and cost trade-offs of each approach.
We’ll also show ZDL in action through a real proof-of-concept use case and explore how Xynomix’s DR services support implementation, testing, monitoring and ongoing recovery readiness.
If you manage Oracle Standard Edition, support customers who do, or want to understand what sub-second RPO on Oracle SE actually means in practice, join us.







