Java, now 32 years old, is already set for its 27th major release. To keep the programming language secure for years and decades to come, Oracle is preparing for the cybersecurity implications of quantum computers. The result is a more secure language, with an added emphasis on productivity gains for programmers.
Three proposals from the Java Development Kit (JDK), known as JDK Enhancement Proposals (JEPs), are designed to make the language quantum-secure. Hybrid key exchange based on TLS 1.3 is intended to be post-quantum secure. When “Q-Day” arrives, the day when all classical cryptographic systems can be cracked by quantum computers, traffic from business-critical applications running on Java version 27 will be secure. Furthermore, PEM encodings of cryptographic objects are intended to ensure that Java integrates better with modern authentication, encryption, and certificate management.
The third quantum-related proposal focuses on diagnosing production systems more securely, that is, without exposing sensitive information. So-called JFR in-process data redaction shields this data, such as customer information or other compliance-driven data.
Keeping pace with the times
The other improvements in Java 27 are focused on ease of use, particularly for AI-related workloads. The Vector API, which serves as a bridge for data traffic in vector format (extremely common in generative AI), enables faster throughput and better performance than was previously possible. Structured concurrency prevents issues with parallel workloads, again, something that is commonplace in AI applications.
AI inference within business logic is less prone to errors in Java 27. Primitive types in patterns, `instanceof`, and `switch` make code clearer in data-driven applications. The latest JEP makes G1 the default garbage collector for all environments, making it easier to achieve predictable performance.
Looking ahead
A preview of JDK 28 EA is Project Valhalla, which is a first step toward narrowing the gap between Java objects and primitives. The latter are more performant but hard-coded into the language, while objects contain pointers and can consume CPU cycles if the relevant data is no longer in cache. In other words: Java needs to become more compact and thus less memory-intensive. Project Leyden is a related initiative, aimed at speeding up Java’s startup time, as well as reducing the time to reach optimal performance and, here too, the amount of memory used.
See also: Oracle releases Java 26: AI, security, and the Java Verified Portfolio