Avaya end of life: the dates that actually bind.

At a glance

Avaya lifecycle date table

Avaya publishes lifecycle notices per release and per commercial model, and the dates that matter are rarely the ones people quote from memory. These are the milestones currently shaping Avaya migration decisions.

Primary sources: Avaya's End of Manufacturer Support notice for Avaya Aura Platform Release 10.1 (notification date 24 May 2024), which also names Aura 10.2 and the planned 10.3 as the on-premises upgrade options, and the Avaya Product Lifecycle Policy. Gateway vintage dates come from Avaya's gateway support notices; entitlements vary by contract and region, so verify yours against your own service agreement.

Read the fine print

What Avaya's lifecycle words actually mean

Three terms do all the work in an Avaya notice, and conflating them is how organizations end up either panicking a year early or discovering they are unsupported a year late.

That third box deserves a second read. Losing access to installers and documentation at obsolescence is a genuine operational risk for an estate you intended to run past support: a hardware failure that requires a reinstall becomes materially harder when the media is no longer downloadable. If you plan to stay on an unsupported release for any length of time, archive your installers, licence files, and documentation now, while they are still available.

It is bigger than one server

What an Aura milestone actually covers

"Avaya Aura" is a platform, not an application. A Release 10.1 support milestone reaches across the components that sit under it, and the practical scope of a migration is nearly always wider than the initial conversation assumes. Depending on your deployment, the estate in scope includes:

• Communication Manager — the call-processing core, plus survivable remote and core servers at branch sites.
• Session Manager and System Manager — SIP routing and the central management plane.
• Session Border Controller for Enterprise (SBCE) — the edge every SIP trunk and remote worker traverses.
• Messaging and Aura Messaging — voicemail, which is frequently the component nobody remembers until cutover.
• Application Enablement Services (AES) — the CTI layer that CRM screen pops, recording, and contact-center integrations depend on.
• Media gateways and media servers — G430, G450, and the analog and PRI interfaces hanging off them.
• Contact center applications — Call Center Elite (now AXP On-Prem), Aura Contact Center, Experience Portal, call recording, and workforce tooling.

Each of those has its own compatibility matrix. A supported Aura release does not automatically make the whole estate supported, and the dependency chain runs both ways: an AES version pinned by your recording platform can block the Aura upgrade you need for gateway support. Map the matrix before you price anything.

The detail that breaks plans

The G430 and G450 vintage trap

This is the Avaya-specific wrinkle that catches otherwise well-run projects, and it has no equivalent in most other vendors' lifecycle notices.

Avaya's gateway support depends on hardware vintage as well as software release. Two G450s that look identical in the rack, run the same firmware, and appear the same in System Manager can sit on opposite sides of a support boundary because they were manufactured years apart. Older G430 and G450 vintages running with Aura 10.2 reach end of manufacturer support on June 30, 2026. Gateways left on Aura 10.1 or earlier stopped receiving firmware and security updates when that release hit EoMS at the start of 2026. Continued long-term support requires both a current Aura release and a recent enough gateway vintage.

The practical step is unglamorous: inventory every gateway with its vintage and serial, cross-reference against the current support matrix, and count how many need replacing. That number frequently changes the economics of the whole decision. If a meaningful share of your gateways require hardware replacement to stay supported, you are no longer comparing "cheap upgrade versus expensive migration" — you are comparing two capital projects, and the cloud option starts looking considerably more reasonable.

After support ends

The practical risk of staying behind

• No security patches. This is the material one. Unified communications platforms sit on the network edge and terminate external SIP traffic; an unpatched SBC or call server is a genuine exposure, not a theoretical one. Your security team will eventually find it on an audit, and the conversation is easier before that happens.
• Support is contract-bound anyway. Even before EoMS, TAC access depends on an active, paid entitlement — and renewal pricing on an aging release rarely improves.
• Compatibility pressure compounds. Certificates expire, browsers change, identity platforms move to modern auth, VMware and hypervisor versions age out, and each of those can break an integration that a patch would have fixed.
• Skills get scarce and expensive. The pool of engineers who can confidently work on an older Aura estate is shrinking, and the ones who remain price accordingly.
• Emergency migrations cost more. A project run under duress after an incident buys the same platform at a worse price with none of the evaluation, because there is no time to compete the quotes.

None of this argues for panic. It argues for a decision with a date on it. Running past EoMS is a legitimate choice when you have a funded plan and a defined window; it is only a problem when it happens by default.

Naming decoder

Call Center Elite, AXP, and what changed

If your Avaya estate includes a contact center, the portfolio renaming is the first thing to untangle, because quotes and documentation now use names that did not exist a few years ago for products you have run for a decade.

AXP Connect is the offer worth understanding properly, because it is designed for exactly the position most Avaya contact centers are in: a large, working, heavily customised on-premises routing estate that cannot be replaced quickly, alongside genuine pressure to add digital channels and AI that the on-premises platform will never deliver. Adding cloud capability on top of existing voice routing is a real option and it defers the hard decision — which is both its strength and the thing to watch. A hybrid that runs for years becomes two platforms to license, administer, and integrate.

The strategic question is simply whether the deferral is buying you something. If it is bridging a defined gap until a planned migration, it is money well spent. If it is a way of avoiding the migration decision indefinitely, you are paying a subscription to keep a lifecycle problem in a holding pattern. Ask what the plan is on the far side of the bridge before you sign.

Decision point

Four realistic paths from an aging Avaya estate

Split employee calling from the contact center

The single most useful structural move available to an Avaya estate: stop treating them as one decision. Employee telephony and customer-service routing have different requirements, different buyers, and different upgrade economics, and there is no technical rule that they must land on the same platform. Plenty of organizations move staff calling to Microsoft Teams Phone or another UCaaS while putting the contact center on a dedicated CCaaS platform. Separating the two usually improves both outcomes and preserves negotiating leverage in each — a vendor bidding for half your estate competes harder than one that assumes it keeps all of it.

The economics

Another on-prem cycle vs a cloud subscription

Upgrading Aura is not a licence purchase; it is another full on-premises cycle. The realistic budget covers software entitlement, gateway replacement where vintages fall out of support, server or virtualization refresh where hosts age out, upgrade services, certificate and firmware work, and regression testing across SBCs, messaging, AES, recording, and the contact center. Recurring costs continue after go-live — maintenance contracts, data-center overhead, specialist Avaya skills that are getting harder to hire — and in five to seven years the cycle repeats.

Cloud converts that into a per-user subscription with no refresh at the end. For contact-center seats, list prices cluster between $65 and $200 per agent per month, with full omnichannel typically $100–$160; employee calling seats price well below agent seats. Our CCaaS pricing guide breaks down the tiers and the add-ons, and the contact center AI guide covers the line item that is growing fastest on every 2026 quote.

The honest comparison is total cost over the same five-to-seven-year window, and it should include two things people leave out. First, the cost of the capability gap: an aging Aura estate does not do digital channels, modern self-service, or AI-assisted quality management, and either you do without them or you buy them separately and integrate. Second, the cost of the next cycle: a paid-off system looks cheap month to month right up until the moment it needs another funded project, at which point the comparison you deferred arrives anyway, with less time to run it.

Work backward from the date

Planning backward from your binding date

Find the earliest date that genuinely constrains you — the Aura EoMS milestone, the June 2026 gateway boundary, a support contract expiry, or an audit finding — and plan in reverse from there. For a mid-sized estate with a contact center attached:

• 3–6 months before: cutover complete. Final porting waves done, parallel run finished, old platform ready to decommission. Porting is the least compressible step — large DID inventories move in waves and rejections reset the clock.
• 6–9 months before: build and pilot. Design signed off, call flows rebuilt and tested, network readiness confirmed, first user and agent groups live on the new platform.
• 9–12 months before: contracts signed. Quotes competed, security review complete, procurement done. This phase reliably consumes a quarter on its own, and security review of a new cloud platform is the step that surprises people.
• 12–18 months before: discovery. Full inventory, gateway vintage audit, requirements, shortlist, and demos against your real call flows rather than the vendor's.

Contact-center dependencies compress the runway further. Elite, Experience Portal, recording, wallboards, and CTI integrations all depend on the Aura layer for call control, so they must move before — or with — the platform underneath them. You cannot retire the call-processing core out from under a live contact center. If your binding date is the June 2026 gateway boundary, the comfortable start has passed; compress discovery deliberately rather than skipping it, and consider a short bridge arrangement to buy planning time rather than rushing the platform decision.

Two shortlists

The Avaya-aligned path vs a vendor-neutral shortlist

The Avaya-aligned path. AXP Public or Private Cloud is the shortest strategic line: familiar terminology, an existing commercial relationship, migration tooling built for exactly your starting point, and admin teams who do not start from zero. Price it first if your estate is deeply Avaya and heavily customised. Validate rather than assume on the specifics — call-flow parity for anything bespoke, CTI and recording integrations, reporting continuity, licensing conversion, and what happens to the customizations that took years to build. And price it against the market regardless: incumbent quotes improve materially when the vendor knows there are credible alternatives on the table.

The vendor-neutral shortlist. A forced migration is the cheapest moment to run a genuine evaluation, because the switching cost is being paid either way. Microsoft-centric organizations should look at Teams Phone with a contact-center layer. Cisco estates in a mixed environment should read the CUCM lifecycle guide and the Webex Contact Center page. For the customer-facing side, dedicated CCaaS platforms such as Five9, Genesys Cloud, and NICE CXone typically out-feature UC-attached routing, and Genesys in particular is a common landing spot for large, complex Avaya contact centers. Healthcare estates should start with the Epic integration guide; public sector buyers with the FedRAMP guide. New to the category? Start with what CCaaS is.

Budget lines

Migration cost factors beyond the subscription

Whichever direction you choose, seat price is only part of the budget. Quotes that look far apart often converge — or flip — once these lines are priced:

• Gateway and hardware replacement. Avaya-specific and frequently the largest surprise: how many G430s and G450s fall below the supported vintage, and what replacing or retiring them costs.
• Porting and carrier work. Moving DIDs, toll-free numbers, and fax lines, plus unwinding legacy PRI or SIP contracts; check for early-termination clauses before you set a date.
• Analog and life-safety. Elevators, alarms, paging, door phones, and fax lines hanging off Avaya gateways. Every one needs a documented replacement, and E911 location requirements need re-establishing on the new platform.
• Contact-center rebuild. Vectors, VDNs, announcements, and reporting do not port. Budget the rebuild as a project, not a task.
• Network readiness. QoS, PoE, wireless coverage, internet redundancy, and firewall changes; some sites need circuit upgrades before voice can move.
• Parallel operation. Weeks or months of running both platforms — double licensing, double administration — while users and queues migrate in waves.
• Training and change management. Rarely itemized in quotes, always paid for in lost productivity if skipped. Agents moving off a hard phone and a familiar desktop need real time.

List prices are a starting point: with competing quotes and term negotiation, mid-market and enterprise deals routinely close 15–30% below list, and the migration line items become negotiable too — implementation credits are often easier to win than rate reductions. Start with a quick estimate of your seat mix.

FAQ

Quick answers

Related guides