Enterprise application modernization services
A bounded upgrade for the application that is blocking your next step.
Compare support and environment changes first. Where application work is needed, agree the scope, price and checks before starting.
Diagnosis when the route is unclear. A code proof when the change and acceptance checks are defined.
Application modernization around one blocker
We'll help remove a concrete blocker to updating an application. First we compare support, an environment update and a bounded code change; then we agree a paid result and how it will be accepted.
Start with the blocker. A component name or an end-of-support date alone does not establish the right project. We check the actual version, support channel, dependencies and the workflow that has to change.
Compare the available paths. Extended support, an environment-only change, your team or provider, and migration tools come first. DFG's role is the remaining work that makes a justified application change deliverable and acceptable. A cheaper sufficient alternative is a valid outcome of the discussion.
Buy one defined result. A paid diagnosis, or a bounded code proof accepted against named workflows. A fixed price follows a defined scope; it is not a promise for an unexamined application portfolio. There is no obligation to continue.
Where applications get stuck
Java application modernization
Java 8, Struts, WebSphere or WebLogic, AIX
The Struts branch gets no more fixes, and the renewal questionnaire still asks about it. The business wants off AIX or WebSphere, and the application holds it there. No team is free for a Struts upgrade that runs for several months.
| Product | Support ends | Source |
|---|---|---|
| Struts 2.5 community support | already passed (2024) | Apache Struts 2.5 end-of-life announcement (opens in a new tab), checked 2026-11-01 |
| Java, WebSphere, AIX | no single date | The vendors publish no single date for these. |
.NET application modernization
.NET Framework on Windows Server 2016 or 2012/R2
A .NET Framework application keeps a Windows Server 2012/R2 or 2016 machine alive near or past the date fixes stop. The external scan before renewal will find that machine. An Azure Arc extension buys time, and someone still has to decide whether the application moves.
| Product | Support ends | Source |
|---|---|---|
| Windows Server 2012/R2 extended updates | already passed (2026-10-13) | Microsoft lifecycle page, checked 2026-11-01 |
| Windows Server 2016 and .NET Framework 4.6.2 | 2027-01-12, extended updates through Azure Arc to 2030 | Windows Server 2016 lifecycle (opens in a new tab), checked 2026-11-01 |
PHP migration and upgrades
PHP up to 8.2 without a distribution backport
A release built from source, a Windows build or a pinned appliance has no backport channel, so nobody ships a fix. Unauthenticated vulnerabilities are announced for releases still in use. Moving to a supported release needs application changes.
| Product | Support ends | Source |
|---|---|---|
| PHP 8.2 | 2026-12-31 | PHP supported versions (opens in a new tab), checked 2026-11-01 |
| PHP 8.3 | 2027-12-31 | PHP supported versions page, checked 2026-11-01 |
<properties>- <java.version>8</java.version>+ <java.version>${supported}</java.version>- <struts.version>2.5</struts.version>+ <struts.version>${supported}</struts.version> </properties> Compilation is one check.
<PropertyGroup>- <TargetFrameworkVersion>v4.6.2</TargetFrameworkVersion>+ <TargetFrameworkVersion>$(Supported)</TargetFrameworkVersion> </PropertyGroup> Compilation is one check.
{ "require": {- "php": "8.2.*"+ "php": "8.3.*" } } Compilation is one check.
This becomes a problem when two things meet: a vulnerability that nobody will fix in your branch, and an outside party that will ask about it on a set day, such as a cyber-insurance renewal questionnaire, an external scan or an auditor. End of support alone is not yet a vulnerability, a deadline or a reason for an outside project.
Start with the blocker
Your team may need to move an application to a different runtime, remove a dependency on an old server environment or establish whether an upgrade is feasible. A component name or an end-of-support date alone does not establish the right project. We first check the actual version, support channel, dependencies and business workflow that needs to change.
- Java
We examine the application together with its runtime, frameworks and deployment environment.
- .NET
We first check whether an environment update or a supported configuration removes the blocker, and only then touch code.
- PHP
We first establish who supports the package and how fixes arrive.
Compare support, environment updates and code changes
Extended support can be a useful bridge when it covers the required risks and time horizon. An environment-only change may remove the blocker without an application rewrite. Your team or current provider may already be able to deliver the necessary change. Migration tools may reduce part of the work; their output still needs an owner and acceptance checks.
DFG's role is the remaining work that makes a justified application change deliverable and acceptable. Compare total customer effort, external cost, operating cost and residual limitations over the same period. A cheaper sufficient alternative is a valid outcome of the discussion.
| Path | Wins when | What we do |
|---|---|---|
| Extended support and backports (for example NES, ESU through Azure Arc) | Coverage is acceptable and cheaper | We recommend it |
| Environment-only update | The application is compatible and the blocker is only the OS or runtime | We check this first |
| Your current contractor or own team | They have the competence, access, time and a way to check the result | We take the bounded remainder or decline |
| Migration tool or accelerator | Its output on your application passes your acceptance checks | We review those results with you; an owner and acceptance checks are still needed |
| Replacing the system | The old system has no product value left | We say so plainly |
A legacy application upgrade keeps the same application and changes its versions, its dependencies and only the code needed to lift the blocker. A migration moves the application to a different runtime, platform or server environment, and a rewrite builds it anew; we do not offer a rewrite by default.
Paid diagnosis or bounded code proof
Buy one defined result. A paid diagnosis, or a bounded code proof accepted against named workflows. A fixed price follows a defined scope; it is not a promise for an unexamined application portfolio. There is no obligation to continue.
Paid diagnosis: a reproducible account of the system, relevant options, critical unknowns and a recommendation for the next decision. The scope names what will be investigated and what will remain outside it. A negative feasibility conclusion can be a properly completed diagnosis.
Bounded code proof: a limited change to one application, with the baseline, intended outcome and critical workflows agreed beforehand. The result can include code, build instructions, regression results, known differences and a handover. The selected component must be relevant to the question being tested; success on an easy fragment does not prove the whole portfolio.
A code proof is not a release permit. It is possible when the change is chosen, there is a way to check the result and the build is reproducible. If nobody can build or check the application, separate paid preparation comes first, or the work does not start.
The price is named after the scope is checked; a scenario calculation is not a price list. Release, data migration, next applications and support are separate decisions and separate purchases.
Paid diagnosis
Bounded code proof
Acceptance evidence and handover
- Name the business workflows that matter, their expected results and the client representative who accepts them.
- Preserve known baseline defects so they are distinguishable from regressions.
- Agree permitted data, environments and access before sharing them.
- Confirm the delivery team and the work needed for release or rollback.
Compilation is one check. It does not establish that the application is ready for production. Production migration, launch support and further applications need their own scope and decision.
Working with code, test data and access
- A master services agreement is available on request.
- Before any access, we agree how the work will be done: in your environment or on suitable test data. If access is not enough, the work does not start.
- Share code and sensitive data only through an agreed protected channel.
- Release and rollback are designed and tested for your particular system; we give no universal downtime guarantee.
Security position
controls mapped to the SOC 2 Trust Services Criteria and HITRUST, for reference, not a certification or audit report
What happens after you ask
The next step is a conversation with the application owner and an available technical specialist.
- What prevents ordinary support or an environment update?
- Who accepts the result and holds the budget?
- Which data, code and checks may be used?
FAQ
Tell us about the application
For the initial request, provide company, work email, role, technology and a short description of the blocker. A decision date is optional. Do not include code, credentials or production data.