# dirtychai

Published articles for dirtychai.

This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles.

## Security Baked Into the JVM: DirtyChai scales vertically, JGDMS scales horizontally

DevFeed: [Security Baked Into the JVM: DirtyChai scales vertically, JGDMS scales horizontally](<https://devfeed.tech/articles/security-baked-into-the-jvm-dirtychai-scales-vertically-jgdms-scales-horizontally-99745.md>)

Original publisher: [Read original article](<https://blog.frankel.ch/security-baked-into-jvm/6/>)

Author: Peter Firmstone

Published: 2026-10-25T00:00:00Z

Content type: article

Language: en

Sources: [Nicolas Fränkel](<https://devfeed.tech/sources/nicolas-frankel.md>)

Topics: [Security](<https://devfeed.tech/topics/security.md>), [Access Control](<https://devfeed.tech/topics/access-control.md>), [least privilege](<https://devfeed.tech/topics/least-privilege.md>), [Process](<https://devfeed.tech/topics/process.md>), [openjdk](<https://devfeed.tech/topics/openjdk.md>)

Tags: [2](<https://devfeed.tech/tags/2.md>), [3](<https://devfeed.tech/tags/3.md>), [allowed](<https://devfeed.tech/tags/allowed.md>), [audit](<https://devfeed.tech/tags/audit.md>), [cache](<https://devfeed.tech/tags/cache.md>), [call](<https://devfeed.tech/tags/call.md>), [capacity](<https://devfeed.tech/tags/capacity.md>), [client](<https://devfeed.tech/tags/client.md>), [code](<https://devfeed.tech/tags/code.md>), [constraint](<https://devfeed.tech/tags/constraint.md>), [dirtychai](<https://devfeed.tech/tags/dirtychai.md>), [java](<https://devfeed.tech/tags/java.md>), [jgdms](<https://devfeed.tech/tags/jgdms.md>), [jini](<https://devfeed.tech/tags/jini.md>), [jvm](<https://devfeed.tech/tags/jvm.md>), [load](<https://devfeed.tech/tags/load.md>), [microservices](<https://devfeed.tech/tags/microservices.md>), [production](<https://devfeed.tech/tags/production.md>), [remote](<https://devfeed.tech/tags/remote.md>), [security](<https://devfeed.tech/tags/security.md>), [series](<https://devfeed.tech/tags/series.md>), [technical](<https://devfeed.tech/tags/technical.md>), [test](<https://devfeed.tech/tags/test.md>), [thread](<https://devfeed.tech/tags/thread.md>)

### AI overview

The article argues that DirtyChai scales within one JVM while JGDMS scales across a fleet, and describes mechanisms intended to keep the security checks practical. DirtyChai gives virtual threads the creator's immutable security context, uses value-based ProtectionDomain identity to reuse cached contexts, and runs policy checks concurrently against immutable grants; reported permission-check costs are about 1.9 µs on platform threads and 4.7 µs on virtual threads under a least-privilege policy. JGDMS adds capacity by registering stateless bytecode-analysis engines, publishes SAFE only after a configured quorum agrees, and keys verdicts by artifact digest. Replicated lookup services and leased registrations support availability; a TTL-bounded cache of signed verdicts can delay revocation during an outage but cannot upgrade a verdict. Authenticated discovery and a bootstrap proxy establish trust and obtain a SAFE verdict before clients unmarshal a full proxy.

### Source excerpt

Nobody runs a SecurityManager in production. The usual reason given is that it does not survive a load test, and every guarantee in this series was bought with a check: a constraint on each call, an audit before each load, a principal set on each thread, a digest on each grant. Part 1 declared constraints on a remote call. Part 2 vetted code before a client loaded it. Part 3 and Part 4 established who is calling and carried that chain across the wire.

## Security Baked Into the JVM: no single party controls the outcome

DevFeed: [Security Baked Into the JVM: no single party controls the outcome](<https://devfeed.tech/articles/security-baked-into-the-jvm-no-single-party-controls-the-outcome-64650.md>)

Original publisher: [Read original article](<https://blog.frankel.ch/security-baked-into-jvm/5/>)

Author: Peter Firmstone

Published: 2026-10-04T00:00:00Z

Content type: article

Language: en

Sources: [Nicolas Fränkel](<https://devfeed.tech/sources/nicolas-frankel.md>)

Topics: [Security](<https://devfeed.tech/topics/security.md>), [DevSecOps](<https://devfeed.tech/topics/devsecops.md>)

Tags: [dirtychai](<https://devfeed.tech/tags/dirtychai.md>), [java](<https://devfeed.tech/tags/java.md>), [jgdms](<https://devfeed.tech/tags/jgdms.md>), [jini](<https://devfeed.tech/tags/jini.md>), [jvm](<https://devfeed.tech/tags/jvm.md>), [microservices](<https://devfeed.tech/tags/microservices.md>), [repo](<https://devfeed.tech/tags/repo.md>), [security](<https://devfeed.tech/tags/security.md>), [technical](<https://devfeed.tech/tags/technical.md>)

### AI overview

The article explains security controls in JGDMS: validating serialized arguments before object creation, granting permissions based on JAR digests and caller identities, and combining policy layers to limit delegated authority. It also describes fail-closed behavior when policies or digest support are unavailable.

### Source excerpt

A URL is a name, and names get reused. Replace the bytes behind https://repo.example.org/order-processor.jar and every permission granted to that URL still applies, now to code nobody audited. So far in this series, Part 1 declared constraints on a remote call, Part 2 vetted code before a client loaded it, Part 3 established who is calling, and Part 4 carried that chain of identities across the wire.

## Security Baked Into the JVM: two Subjects, one call

DevFeed: [Security Baked Into the JVM: two Subjects, one call](<https://devfeed.tech/articles/security-baked-into-the-jvm-two-subjects-one-call-18930.md>)

Original publisher: [Read original article](<https://blog.frankel.ch/security-baked-into-jvm/3/>)

Author: Peter Firmstone

Published: 2026-08-09T00:00:00Z

Content type: article

Language: en

Sources: [Nicolas Fränkel](<https://devfeed.tech/sources/nicolas-frankel.md>)

Topics: [Security](<https://devfeed.tech/topics/security.md>), [Next-Gen Identity Security](<https://devfeed.tech/topics/next-gen-identity-security.md>)

Tags: [api](<https://devfeed.tech/tags/api.md>), [bearer](<https://devfeed.tech/tags/bearer.md>), [call](<https://devfeed.tech/tags/call.md>), [certificates](<https://devfeed.tech/tags/certificates.md>), [client](<https://devfeed.tech/tags/client.md>), [code](<https://devfeed.tech/tags/code.md>), [constraint](<https://devfeed.tech/tags/constraint.md>), [credentials](<https://devfeed.tech/tags/credentials.md>), [dirtychai](<https://devfeed.tech/tags/dirtychai.md>), [dispatch](<https://devfeed.tech/tags/dispatch.md>), [filter](<https://devfeed.tech/tags/filter.md>), [fleet](<https://devfeed.tech/tags/fleet.md>), [frameworks](<https://devfeed.tech/tags/frameworks.md>), [identity](<https://devfeed.tech/tags/identity.md>), [java](<https://devfeed.tech/tags/java.md>), [jgdms](<https://devfeed.tech/tags/jgdms.md>), [jini](<https://devfeed.tech/tags/jini.md>), [jvm](<https://devfeed.tech/tags/jvm.md>), [microservices](<https://devfeed.tech/tags/microservices.md>), [pipeline](<https://devfeed.tech/tags/pipeline.md>), [security](<https://devfeed.tech/tags/security.md>), [technical](<https://devfeed.tech/tags/technical.md>), [thread](<https://devfeed.tech/tags/thread.md>), [token](<https://devfeed.tech/tags/token.md>), [verify](<https://devfeed.tech/tags/verify.md>)

### AI overview

DirtyChai separates process workload identity from human user identity. A JVM-wide process Subject carries a SPIFFE SVID and is available across requests; a per-request user Subject carries identities established through JWT/OIDC, with Kerberos retained for legacy deployments. The excerpt says SPIRE issues short-lived X.509 SVIDs through a local agent, avoiding long-lived keystores, and that DirtyChai is replacing legacy JAAS subject handling with explicit identity and privilege boundaries. It also warns that JAAS doAsPrivileged can shed the enclosing user identity as a side effect. The supplied excerpt ends during the SPIRE component description, so later implementation details are not available.

### Source excerpt

The constraint system stops a bad call before it leaves the JVM. The Safe Codebase Audit Pipeline stops bad code before a client ever loads it. What remains is identity: who is calling, and can you verify it? Most frameworks answer with a token check at the door. A filter validates a bearer token, sets a thread-local variable, and hopes that nothing downstream forgets to look at it. DirtyChai answers differently.

## Security Baked Into the JVM: the Safe Codebase Audit Pipeline

DevFeed: [Security Baked Into the JVM: the Safe Codebase Audit Pipeline](<https://devfeed.tech/articles/security-baked-into-the-jvm-the-safe-codebase-audit-pipeline-18929.md>)

Original publisher: [Read original article](<https://blog.frankel.ch/security-baked-into-jvm/2/>)

Author: Peter Firmstone

Published: 2026-07-19T00:00:00Z

Content type: article

Language: en

Sources: [Nicolas Fränkel](<https://devfeed.tech/sources/nicolas-frankel.md>)

Topics: [software supply-chain attack](<https://devfeed.tech/topics/software-supply-chain-attack.md>), [code security](<https://devfeed.tech/topics/code-security.md>), [software-architecture](<https://devfeed.tech/topics/software-architecture.md>), [Data serialization](<https://devfeed.tech/topics/data-serialization.md>)

Tags: [audit](<https://devfeed.tech/tags/audit.md>), [bytecode](<https://devfeed.tech/tags/bytecode.md>), [dirtychai](<https://devfeed.tech/tags/dirtychai.md>), [distributed](<https://devfeed.tech/tags/distributed.md>), [java](<https://devfeed.tech/tags/java.md>), [jgdms](<https://devfeed.tech/tags/jgdms.md>), [jini](<https://devfeed.tech/tags/jini.md>), [jvm](<https://devfeed.tech/tags/jvm.md>), [malicious](<https://devfeed.tech/tags/malicious.md>), [microservices](<https://devfeed.tech/tags/microservices.md>), [pipeline](<https://devfeed.tech/tags/pipeline.md>), [security](<https://devfeed.tech/tags/security.md>), [supply-chain](<https://devfeed.tech/tags/supply-chain.md>), [supply-chain-compromise](<https://devfeed.tech/tags/supply-chain-compromise.md>), [technical](<https://devfeed.tech/tags/technical.md>)

### AI overview

The excerpt presents SCAP as a way to inspect third-party JARs before clients deserialize objects from them, addressing the risk that runtime-loaded code may be replaced with malicious bytecode. It identifies two targets: dangerous bytecode that can arrive through legitimate permission grants, and blocking static initializers that can stall virtual-thread scheduling. SCAP uses SHA-256 content hashes so a JAR is analyzed once regardless of which services serve it. The described architecture assigns separate roles to five hosts and limits network access; the excerpt specifies that the Lookup Service stores service items opaquely and emits registration events, while the Codebase Downloader alone has outbound internet access and proactively fetches JARs on those events. The excerpt cuts off before describing the remaining hosts.

### Source excerpt

In Part 1, the minimal deployment showed constraints traveling with the proxy: authentication, encryption, hardened deserialization, all declared in configuration and enforced at the call boundary. The proxy is a JAR. That JAR was downloaded and unmarshalled before any constraint ran. That step is the earlier problem. Distributed Java systems that load remote code are vulnerable to supply chain compromise: an attacker can replace a legitimate JAR with one containing malicious bytecode.

## Security Baked Into the JVM: why fork Apache River and OpenJDK?

DevFeed: [Security Baked Into the JVM: why fork Apache River and OpenJDK?](<https://devfeed.tech/articles/security-baked-into-the-jvm-why-fork-apache-river-and-openjdk-18928.md>)

Original publisher: [Read original article](<https://blog.frankel.ch/security-baked-into-jvm/1/>)

Author: Peter Firmstone

Published: 2026-06-28T00:00:00Z

Content type: article

Language: en

Sources: [Nicolas Fränkel](<https://devfeed.tech/sources/nicolas-frankel.md>)

Topics: [Security](<https://devfeed.tech/topics/security.md>), [Authorization](<https://devfeed.tech/topics/authorization.md>), [openjdk](<https://devfeed.tech/topics/openjdk.md>), [jini](<https://devfeed.tech/topics/jini.md>), [software-architecture](<https://devfeed.tech/topics/software-architecture.md>), [trust](<https://devfeed.tech/topics/trust.md>), [Cryptography](<https://devfeed.tech/topics/cryptography.md>), [Data serialization](<https://devfeed.tech/topics/data-serialization.md>), [Network](<https://devfeed.tech/topics/network.md>), [Deployment](<https://devfeed.tech/topics/deployment.md>)

Tags: [apache](<https://devfeed.tech/tags/apache.md>), [auditing](<https://devfeed.tech/tags/auditing.md>), [authorization](<https://devfeed.tech/tags/authorization.md>), [bytecode](<https://devfeed.tech/tags/bytecode.md>), [code](<https://devfeed.tech/tags/code.md>), [deprecated](<https://devfeed.tech/tags/deprecated.md>), [dirtychai](<https://devfeed.tech/tags/dirtychai.md>), [distributed](<https://devfeed.tech/tags/distributed.md>), [firewall](<https://devfeed.tech/tags/firewall.md>), [fork](<https://devfeed.tech/tags/fork.md>), [inside](<https://devfeed.tech/tags/inside.md>), [java](<https://devfeed.tech/tags/java.md>), [java-17](<https://devfeed.tech/tags/java-17.md>), [jgdms](<https://devfeed.tech/tags/jgdms.md>), [jini](<https://devfeed.tech/tags/jini.md>), [jvm](<https://devfeed.tech/tags/jvm.md>), [microservices](<https://devfeed.tech/tags/microservices.md>), [network](<https://devfeed.tech/tags/network.md>), [network-firewall](<https://devfeed.tech/tags/network-firewall.md>), [openjdk](<https://devfeed.tech/tags/openjdk.md>), [security](<https://devfeed.tech/tags/security.md>), [technical](<https://devfeed.tech/tags/technical.md>), [third-party](<https://devfeed.tech/tags/third-party.md>), [trust](<https://devfeed.tech/tags/trust.md>), [trust-boundaries](<https://devfeed.tech/tags/trust-boundaries.md>)

### AI overview

The excerpt presents DirtyChai and JGDMS as separate projects designed to work together after Java's SecurityManager removal left applications that load remote code without built-in JVM authorization controls. DirtyChai is a community OpenJDK fork that restores authorization infrastructure and is required to run JGDMS; it adds virtual-thread support and a lock-free policy engine within one JVM. JGDMS is a security-hardened Apache River fork for dynamically discovering microservices over IPv6. It uses lease-based Jini-style discovery, JERI for pluggable constraint-based RPC, and layered protections including hardened deserialization, TLS 1.3, authorization, proxy trust checks, and bytecode analysis before loading. The excerpt introduces a deployment example but does not provide its steps.

### Source excerpt

The more distributed a system, the harder it is to secure. Code crosses JVM boundaries. Objects are serialized across trust boundaries. Third-party proxies run inside your process. The usual answer is a network firewall. It helps, but it operates at the wrong level. Java 17 deprecated the SecurityManager, Java 24 put the final nail in its coffin. Most developers didn't notice.