If your Exchange Server ESU budget is not approved, the right question is not whether every mailbox can move this quarter. It is which mail-related workloads should leave first so one exposed server does not keep carrying every user, every device, every archive dependency and every internet-facing protocol. That matters more now because Australian reporting in September 2026 found 382 Australian Exchange servers still vulnerable to CVE-2026-62911 on 31 August, three weeks after Microsoft's fix, with public exploit code available. Microsoft has also been explicit that Exchange 2016 and 2019 support ended on 14 October 2025, and that the paid Exchange 2016 and 2019 ESU Period 2 ends when October 2026 ends, with no further extension. For Australian organisations that still hold personal information in mail systems, the pressure is not only operational. The OAIC says APP 11 requires entities covered by the Privacy Act to take active measures to ensure the security of personal information they hold. So if a full exit is unaffordable right now, treat Exchange as a workload triage exercise. Move the workloads that most reduce attack surface, credential exposure and recovery complexity first, then leave the awkward edge cases for a controlled second phase.
01Start with workload triage, not with org charts
Many mid-market teams start by asking which business unit can migrate first. That sounds sensible, but it is usually the wrong cut if your problem is exposure. A safer first pass is by workload type: interactive user mailboxes, shared and service mailboxes, scanners and line-of-business senders, public folders, journaling and archive integrations, and any admin or hybrid dependencies. Each group has a different risk profile and a different amount of effort to move.
The triage goal is simple: remove the workloads that force your old Exchange estate to stay internet facing, credential heavy or operationally central. That means your first moves are rarely the hardest moves. They are the ones that reduce how many people and systems still depend on the box. If you are still deciding whether to hold on to Exchange at all, this is also where a broader legacy modernisation lens helps, because the mail server often survives only because a few adjacent processes never got redesigned.
Use this rule of thumb. Prioritise workloads in this order: internet-facing user access, privileged admin access, business-user shared mailboxes, scanner and application senders, old public folders, then compliance edge cases such as journaling destinations and specialist archives. The order is driven by how much each category shrinks the blast radius, not by how politically easy the migration is.
- Ask which workloads keep Exchange exposed to the internet
- Ask which workloads still rely on passwords or legacy auth patterns
- Ask which workloads would multiply the impact of a compromise
- Ask which workloads can move with native Microsoft methods
- Ask which workloads are blocked by archive, legal hold or line-of-business dependencies
02Leave first: interactive user mailboxes that still touch the old server
User mailboxes usually deliver the biggest reduction in risk per week of effort. In a hybrid deployment, Microsoft supports moving mailboxes between on-premises Exchange and Exchange Online using migration batches and the Exchange Admin Center. Once users are off the old server, the remaining estate becomes a smaller technical island rather than the centre of daily communication.
Why these first? Because user mailboxes drive the most authentication traffic, the most browser exposure through Outlook on the web and admin portals, the most helpdesk load, and the largest amount of personal information sitting on a system you may soon be unable to patch. For organisations using Microsoft 365 identity controls, Microsoft also documents Hybrid Modern Authentication so on-premises mailboxes can use cloud-issued tokens and stronger authentication patterns. But if your budget is constrained, enabling HMA just to keep a shrinking on-prem estate alive is often a bridging move, not the end state.
In practice, move standard knowledge-worker mailboxes before you worry about every edge case. Keep highly customised executive assistants, litigation hold exceptions or unusual delegate workflows for the second wave if needed. The point is to cut the bulk of daily user dependency quickly. If you are also rethinking long term mail architecture, our team at SyncTrix often sees the best outcomes when this phase is run as a mail programme, not just an infrastructure ticket queue, because mailbox moves trigger identity, support, device and communications changes at the same time.
- Move standard user mailboxes in batches
- Keep unusual delegate and compliance exceptions for a later wave
- Retest mobile, Outlook profile and shared mailbox access after each batch
- Use the mailbox move to reduce OWA and ECP dependence on the old estate
03Second: shared mailboxes, team inboxes and low-complexity service mailboxes
After user mailboxes, shared mailboxes and team inboxes are often the best next target. They usually matter to operations, but their migration complexity is lower than full user populations. They also remove a lot of daily traffic from the old environment. Exchange Online has documented service limits for mailboxes, and Microsoft also documents how to manage resource mailboxes such as room and equipment mailboxes.
This category includes finance inboxes, support queues, HR team addresses, no-reply mailboxes that are still manually monitored, and basic room or equipment mailboxes. If they are under sane size and permission structures, they are usually quick wins. They also reduce the number of people who still need to open Outlook against the old environment just to check a team inbox. For businesses already looking at mailbox cost and account sprawl, this pairs well with our earlier post on which shared mailboxes and service accounts should leave first.
Be more careful with so-called service mailboxes that are actually being used as application identities. If a mailbox is tied to a device, workflow engine or legacy SMTP process, do not treat it as an ordinary shared mailbox. Split the ownership question first: is it a collaboration inbox or an integration credential? If it is the latter, push it into the application-sender workstream below.
- Move ordinary shared mailboxes early
- Move room and equipment mailboxes if they are not tied to old booking add-ins
- Separate collaboration inboxes from application identities before migration
- Rationalise permissions while usage is visible
| Workload | Why it should move now | Typical blockers | Priority |
|---|---|---|---|
| Standard user mailboxes | Cuts daily user dependence, web exposure and authentication traffic | Delegate quirks, mobile profile issues, legal hold exceptions | 1 |
| Shared and team mailboxes | Removes high-traffic operational inboxes from old Exchange | Messy permissions, oversized mailboxes, hidden app use | 2 |
| Scanners, printers and app senders | Lets you isolate or remove risky credential-based sending paths | Old devices, SMTP AUTH dependence, relay design | 3 |
| Public folders | Reduces long-tail dependence on old collaboration structures | Migration downtime, mapping issues, ownership gaps | 4 |
| Journaling and archive dependencies | Important for final exit, but often not the fastest blast-radius reduction | Third-party archive design, compliance sign-off | 5 |
04Third: scanners, printers, line-of-business apps and other send-only workloads
This is the category many teams leave too late. That is a mistake. Device and application mail often carries ugly credential habits, broad relay permissions and weak visibility. Microsoft documents four main ways to send mail from devices or apps via Microsoft 365: Client SMTP submission, SMTP relay, Direct Send and High Volume Email. Microsoft also recommends you disable SMTP AUTH organisation-wide and only enable it for accounts that still require it.
That recommendation matters because scanners and legacy applications are often the last reason an old Exchange box remains central to message flow. If you can move these senders to Exchange Online relay, a cloud-based mail service, or a redesigned application notification path, you cut a disproportionate amount of risk. Newer patterns are clearer too. Microsoft now documents High Volume Email for Microsoft 365 for internal automated and operational messaging, and also documents connectors for relaying mail from devices and applications.
Treat each sender as one of three things: can modernise now, can relay safely with controls, or must be retired. A scanner that only emails internal users may fit High Volume Email. A line-of-business app sending external customer notifications may need a relay connector or a non-Exchange mail service. A device that only supports poor authentication may force a replacement decision. This is where a good mail infrastructure or cloud infrastructure team can save time, because the technical work is not hard individually, but there are usually dozens of hidden senders.
- Inventory every SMTP sender before mailbox migration closes
- Classify each sender as internal-only, external transactional, or user-delegated
- Prefer relay or dedicated service patterns over human mailbox credentials
- Replace devices that cannot meet your authentication and logging requirements
- Turn off tenant-wide SMTP AUTH if only a small exception set remains
05Fourth: public folders and old collaboration leftovers
Public folders are rarely the first security problem, but they often become the reason the old platform cannot fully die. Microsoft provides a supported way to batch migrate Exchange Server public folders to Microsoft 365, and also documents hybrid access to on-premises public folders from Exchange Online users.
Why not do this first? Because public folders usually create less direct internet-facing risk than live user access and app sending. Why not leave them until the very end? Because if they remain, they tend to keep old admin knowledge, backup processes and support runbooks alive. They also usually hide ownership problems. Nobody quite knows which team still depends on them until cutover is close.
For mid-market organisations, the realistic move is often: stabilise user and send-only workloads first, then either migrate public folders properly or replace them with a simpler target such as shared mailboxes, SharePoint or a line-of-business system. If your collaboration design is already changing, this sits naturally with broader digital engineering and API integration work, because the folder itself is often just a symptom of an old business process.
- Use public folder usage analysis before promising a migration date
- Expect some downtime for final migration steps
- Challenge whether each public folder should exist at all in the target state
- Move active business content, not historical clutter by default
06Leave later, but design early: journaling, retention and archive dependencies
Compliance workloads are not always the first thing to move, but they must be understood early because they can block final decommissioning. Microsoft states that Exchange Online does not support delivering journal reports to an Exchange Online mailbox. If journaling is still part of your design, the destination must be an on-premises archiving system or a third-party archiving service. Microsoft also says that when possible it recommends Microsoft 365 retention for in-place archiving and management rather than journaling for every scenario.
This matters because some organisations assume they can move user mailboxes first and work out archive design later. Sometimes that is fine. Sometimes it strands an old Exchange server purely because a legal or archive feed still depends on it. The answer is not to make compliance the first migration wave. The answer is to get legal, records and security stakeholders into the design now, before your mailbox moves create a dead end.
Australian organisations should also remember that the OAIC's APP 11 guidance is about taking reasonable steps to secure personal information and actively considering retention. That does not tell you which archive product to buy, but it does mean old mail stores should not survive indefinitely just because no one owns the decision. If you are simultaneously reviewing mailbox retention and dormant accounts, see our post on whether archived project mail should move off licensed user accounts first.
- Document every journaling and archive dependency before final cutover planning
- Check whether retention can replace old journaling assumptions
- Get legal and records teams to sign off before decommission dates are committed
- Do not let archive uncertainty delay user and sender risk reduction
07What to ask a migration partner when funding only covers a staged exit
If you cannot fund a complete programme now, the partner you choose needs to think in risk-reduction stages, not just mailbox counts. Ask them how they would reduce internet exposure in the first 30 days, which workloads they would move without waiting for every exception, and what controls they would put around anything left behind. The ASD's Information Security Manual, September 2026 release is a reminder that Australian organisations need a clear control posture around systems exposed to cyber threats.
A strong answer should include workload inventory, migration waves, fallback plans, send-only workload remediation, archive dependency mapping, and a target point at which the old Exchange server can be ring-fenced or removed from internet-facing duty. If the answer is only 'we will migrate users in batches', you are buying execution without triage.
Useful vendor questions are specific. Which remaining workloads would still require Exchange to stay internet facing after wave one? How will you identify application and device senders that do not appear in ordinary mailbox reports? What is your rollback method for migration batches? What evidence will you provide that old protocols and relay paths were actually retired? Those are the questions that expose whether the team can shrink blast radius, not just move data. If you need that kind of programme support, SyncTrix provides this through its mail infrastructure service.
- How do you rank workloads by blast-radius reduction, not just migration ease?
- Which protocols and public endpoints can be removed after wave one?
- How will you inventory hidden SMTP senders and relay dependencies?
- How will you handle journaling, archives and legal hold edge cases?
- What proof will you give that the old estate is no longer carrying user traffic?
Sources
- Hundreds of old, vulnerable Exchange servers remain in Australia
- Reminder: Exchange 2016 and 2019 ESU Program Ends in October 2026
- Exchange Server 2019 and 2016 end of support roadmap
- Chapter 11: APP 11 Security of personal information
- How to configure Exchange Server on-premises to use Hybrid Modern Authentication
- Move mailboxes between on-premises and Exchange Online organizations in hybrid deployments
- Use the EAC to move mailboxes
- How to set up a multifunction device or application to send email using Microsoft 365 or Office 365
- Enable or disable SMTP AUTH in Exchange Online
- Manage High Volume Email for Microsoft 365
- Configure mail flow using connectors in Exchange Online
- Batch migrate Exchange Server public folders to Microsoft 365
- Configure Exchange Server public folders for a hybrid deployment
- Configure Journaling in Exchange Online
- Journaling in Exchange Online
- Exchange Online limits
- Manage resource mailboxes in Exchange Online
Topics
SyncTrix Editorial
SyncTrix · SyncTrix
Writes about the engineering decisions behind production systems - architecture, delivery and the trade-offs that only show up at scale.
Building something like this?
SyncTrix engineers AI, SaaS, platform and cloud systems for enterprises and high-growth teams. Tell us what you're shipping and we'll scope it with you.