# Ztunnel

This document provides an overview of the architecture and design decisions around Ztunnel, the node-proxy component in ambient mode.

## Background and motivation

Motivations to implement ztunnel generally came from two areas.

First, and most importantly, it serves as a means to implement the real goal: waypoints.

        notActivated.setId("notActivated");

        Activation nonActivation = new Activation();

        nonActivation.setJdk("19.2");

        notActivated.setActivation(nonActivation);

        Profile defaultActivated = new Profile();
        defaultActivated.setId("defaultActivated");

        Activation defaultActivation = new Activation();

        defaultActivation.setActiveByDefault(true);

            return builder;
        }

        @Override
        protected Activation.Builder transformActivation_File(
                Supplier<? extends Activation.Builder> creator, Activation.Builder builder, Activation target) {
            stk.push(nextFrame("file", Activation::getFile));
            Optional.ofNullable(target.getFile());

        boolean modified = false;
        for (Profile profile : model.getProfiles()) {
            Activation activation = profile.getActivation();
            if (activation != null) {
                // restore activation
                profile = profile.withActivation(activations.get(profile.getId()));
                modified = true;
            }
            profiles.add(profile);
        }

        </executions>
      </plugin>
    </plugins>
  </build>

  <profiles>
    <profile>
      <id>create-distribution-in-dir</id>
      <activation>
        <property>
          <name>distributionTargetDir</name>
        </property>
      </activation>
      <build>
        <plugins>
          <plugin>
            <groupId>org.apache.maven.plugins</groupId>

*   Add `UnifiedGRU` as the new GRU implementation for tf2.0. Change the default
    recurrent activation function for GRU from `hard_sigmoid` to `sigmoid`, and
    `reset_after` to True in 2.0. Historically recurrent activation is
    `hard_sigmoid` since it is fast than 'sigmoid'. With new unified backend
    between CPU and GPU mode, since the CuDNN kernel is using sigmoid, we change

culture and the further production of creative, cultural and scientific
works, or to gain reputation or greater distribution for their Work in
part through the use and efforts of others.

For these and/or other purposes and motivations, and without any
expectation of additional consideration or compensation, the person
associating CC0 with a Work (the "Affirmer"), to the extent that he or she