Since Oracle deprecated Java Web Start in 2017, most organisations have picked a quick fix. Many used OpenWebStart to restore JNLP on the desktop. It kept operations running. The decision now is about the next three years: keep a desktop launcher and a local OpenJDK installation, or move to a browser-native runtime. For organisations looking for an OpenWebStart alternative, the strategic choice is whether to keep a desktop launcher and local OpenJDK installation or move Java applications to a browser-native runtime.
Parity is easy. Progress is harder.
Java Web Start is familiar. Double-click a JNLP and the app opens on the desktop. The trouble is the quiet costs that come with it: keeping a maintained local Java installation, version update prompts, version drift and a wider audit surface from local JREs, especially for OpenJDK distributions. None of these are deal breakers on day one, but they become expensive at fleet scale.
CheerpJ takes a different path. It runs unmodified Java inside the browser. Users open a JNLP file, and the application starts on the page, without ever leaving the browser. No plugins or a local Java installation. For estates trying to simplify the endpoint and standardise on browser security controls, that change matters.
What changes when you choose the browser
Endpoint footprint shrinks. There is no client to install, no JRE to patch, no version drift to chase. IT shifts effort from packaging to publishing.
Security posture improves. You lean on SSO, CSP and standard headers, and you remove unmanaged Java from endpoints. The browser session executes client-side code and calls your existing services.
User experience gets simpler. Launch from a portal, bookmark a URL, pick up where you left off. First load is cached, warm starts are quick, printing and downloads use familiar browser flows.
Operations look like the rest of the web. You host JARs next to the runtime, push changes once, and measure real user performance the same way you do for other web apps.
Case study: CF Fertilisers
Context
CF Fertilisers needed to keep Oracle E‑Business Suite (R12.1.13) Forms available while removing Java 8u311 and IE mode from user desktops in the US and UK. The objective was to improve their Rapid7 security posture and eliminate the local Java dependency without disrupting operations.
Approach
They adopted CheerpJ to deliver the EBS Forms client in the browser. No plugins and no local JRE. Access is launched from a controlled storefront that is also used by UK group policy.
Outcome
Desktop Java was removed, EBS Forms runs in modern browsers, and the security posture improved by retiring the old JRE. Support load dropped because users no longer see Java version prompts or IE mode issues, and change management now happens at the web tier.
Where each option fits
OpenWebStart is a short-term bridge when you need desktop JNLP parity and have strong device management in place. It is not a browser solution and it keeps local Java on endpoints.
CheerpJ is a browser-native model for enterprises that want to run an OpenJDK app in a browser without endpoint installs. Core runs on Chrome, Edge, Firefox and Safari. The JNLP Runner keeps your descriptors for Chromium browsers. You keep your server-side stack and move access to the web.
The strategic view in one table
| Dimension | OpenWebStart | CheerpJ |
| Endpoint installs | Yes | No |
| Runs in the browser | No | Yes |
| Security and audit | Larger scope with local JREs | Smaller scope with web controls |
| Ongoing effort | Device packaging and updates | Standard web hosting and change control |
| Cross-browser coverage | Desktop only | Chrome, Edge, Firefox, Safari |
| Fit for mobile and mixed fleets | Poor | Good |
What about Oracle estates
If you run Oracle Forms or EBS, CheerpJ provides a proven route to the browser while leaving the server-side intact. Many organisations kept access alive with OpenWebStart or VDI. Moving to CheerpJ removes desktop Java and gives users direct browser access instead.
The call to make
If your priority is short-term continuity on managed desktops, OpenWebStart delivers. If your priority is a durable model that reduces endpoint risk, improves user access and aligns with how the rest of your estate runs, choose CheerpJ.
Bottom line: to run OpenJDK apps in browsers at scale, CheerpJ is the long-term strategy.
Want to prove it
Pick one workflow. Publish the JARs, add the bootstrap page or keep your JNLP with the CheerpJ JNLP Runner, and point a small group of users to the browser route. Measure time to launch and helpdesk volume against your current setup. Scale from there.
