DefaultClusterLocalDomain = "cluster.local"

	// DefaultClusterSetLocalDomain is the default domain suffix for Kubernetes Multi-Cluster Services (MCS)
	// used for load balancing requests against endpoints across the ClusterSet (i.e. mesh).
	DefaultClusterSetLocalDomain = "clusterset.local"

	// DefaultClusterName is the default cluster name
	DefaultClusterName = "Kubernetes"

	"istio.io/istio/pkg/xds"
)

var (
	log = xds.Log

	// Tracks connections, increment on each new connection.
	connectionNumber = int64(0)
)

// Used only when running in KNative, to handle the load balancing behavior.
var firstRequest = uatomic.NewBool(true)

var knativeEnv = env.Register("K_REVISION", "",
	"KNative revision, set if running in knative").Get()

// DiscoveryStream is a server interface for XDS.

			"all traffic on that port will be matched, which can have disastrous consequences. Additionally, the destination is seen as an opaque destination; "+
			"even if it is another service in the mesh, policies such as mTLS and load balancing will not be used when connecting to it.").Get()

	ValidateWorkloadEntryIdentity = env.Register("ISTIO_WORKLOAD_ENTRY_VALIDATE_IDENTITY", true,
		"If enabled, will validate the identity of a workload matches the identity of the "+

// a particular source instance, as there should typically be one instance
// per cluster for any deployment. However we do not identify a destination
// cluster, as we expect most tests will cause load-balancing across all possible
// clusters.
func (t *T) Run(testFn oneToOneTest) {
	t.rootCtx.Logf("Running tests with: sources %v -> destinations %v",
		t.sources.Services().NamespacedNames().NamesWithNamespacePrefix(),

	// Resolution indicates how the service instances need to be resolved before routing
	// traffic. Most services in the service registry will use static load balancing wherein
	// the proxy will decide the service instance that will receive the traffic. Service entries
	// could either use DNS load balancing (i.e. proxy will query DNS server for the IP of the service)

  repeated string subject_alt_names = 6;
  // Waypoint is the waypoint proxy for this service. When set, all incoming requests must go
  // through the waypoint.
  GatewayAddress waypoint = 7;

  // Load balancing policy for selecting endpoints.
  // Note: this applies only to connecting directly to the workload; when waypoints are used, the waypoint's load_balancing
  // configuration is used.
  LoadBalancing load_balancing = 8;
}

//
// For cases where all clusters are on the same network, verifies that each of the expected clusters was reached.
//
// For multi-network configurations, verifies the current (limited) Istio load balancing behavior when going through
// a gateway. Ensures that all expected networks were reached, and that all clusters on the same network as the
// client were reached.

					toMatch:    match.And(match.NotVM, notMigration),
					expectMTLS: notNaked,
					// Since we are doing pass-through, all requests will stay in the same cluster,
					// as we are bypassing Istio load balancing.
					// TODO(https://github.com/istio/istio/issues/39700): Why does headless behave differently?
					expectCrossCluster: and(notFromNaked, or(toHeadless, toStatefulSet)),
					expectCrossNetwork: never,

	if cfg.Extra.ServiceNodePortRange.Size == 0 {
		// TODO: Currently no way to specify an empty range (do we need to allow this?)
		// We should probably allow this for clouds that don't require NodePort to do load-balancing (GCE)
		// but then that breaks the strict nestedness of ServiceType.
		// Review post-v1
		cfg.Extra.ServiceNodePortRange = kubeoptions.DefaultServiceNodePortRange

Diese verteilten Containerverwaltungssysteme wie Kubernetes verfügen normalerweise über eine integrierte Möglichkeit, die **Replikation von Containern** zu handhaben und gleichzeitig **Load Balancing** für die eingehenden Requests zu unterstützen. Alles auf **Cluster-Ebene**.