Compendium appendix d feature status

Navigation: Home > Pages

Appendix D: Feature Status Matrix

Version: 1.8.0
Stand: April 2026
Kategorie: Feature-Übersicht und Implementierungsstatus

---

Zweck dieses Appendix

Dieser Appendix bietet eine vollständige, strukturierte Übersicht aller ThemisDB-Features mit aktuellem Implementierungsstatus, Dokumentationsverweisen und Roadmap-Planung. Dies dient als zentraler Referenzpunkt für:

• Entwickler: Welche Features sind verfügbar? Wo ist der Code?
• Architekten: Welche Capabilities hat ThemisDB? Was fehlt noch?
• Projektmanager: Was ist Production-Ready? Was ist geplant?
• Nutzer: Welche Funktionen kann ich nutzen?

---

Status-Definitionen

Symbol	Status	Bedeutung
✅	Production-Ready	Vollständig implementiert, getestet (>85% Coverage), dokumentiert, performance-validated
🟢	MVP Complete	Kernfunktionalität vorhanden, stabil, aber ggf. noch Optimierungen ausstehend
🟡	In Development	Aktiv in Entwicklung, teilweise funktional, noch nicht production-grade
🟠	Design Phase	Spezifikation vorhanden, Implementierung geplant
🔵	Planned	Auf Roadmap, noch keine Spezifikation
⚪	Backlog	Idee/Anforderung dokumentiert, keine aktive Planung
🔴	Alpha	Frühes Stadium, API kann sich ändern
❌	Not Planned	Bewusst nicht im Scope

---

Zusammenfassung (v1.8.0, April 2026)

Bereich	Status
Core Database Engine	✅ Production-Ready
Multi-Model (Relational/Graph/Vector/Document/TS)	✅ Production-Ready
ACID Transactions / MVCC / SAGA	✅ Production-Ready
Security & Compliance (GDPR/HIPAA/CCPA/PCI/SOC2)	✅ Production-Ready
Horizontal Scaling / Sharding	✅ Production-Ready
High Availability / Replication	✅ Production-Ready
LLM/RAG Integration (llama.cpp)	✅ Production-Ready
Full-Text + Vector Hybrid Search	✅ Production-Ready
Analytics (OLAP/CEP/Process Mining)	✅ Production-Ready
Observability (Prometheus/OTel/Grafana)	✅ Production-Ready
Voice Assistant	🟢 MVP Complete
Chimera Multi-DB Adapter	🔴 Alpha
GPU Compute (CUDA/Vulkan/OpenCL)	🟢 MVP Complete
PostgreSQL Wire Protocol	🟢 MVP Complete

Overall Production-Ready: 40/46 modules (87%)

---

v1.8.0 Feature-Highlights

Authentication & Authorization (§21)

Feature	Status	Class	Documentation
JWT Validator (scope enforcement)	✅	JWTValidator	§21.1
OAuth2 PKCE	✅	OAuth2PKCEHandler	§21.2
SAML 2.0	✅	SAMLAuthenticator	§21.3
WebAuthn / FIDO2	✅	WebAuthnAuthenticator	§21.4
LDAP / Active Directory	✅	LDAPAuthenticator	§21.5
MFA (TOTP/SMS/Email)	✅	MFAManager	§21.6
Session Management	✅	SessionManager	§21.7
Password Policy Engine	✅	PasswordPolicyEngine	§21.8
Federated OIDC	✅	FederatedOIDCProvider	§21.9
GSSAPI / Kerberos	✅	GSSAPIAuthenticator	§21.10

Module: src/auth/, include/auth/
Compendium: chapter_21_auth.md

---

Bi-Temporal & Time-Series (§9)

Feature	Status	Class	Documentation
BiTemporalTable (valid + transaction time)	✅	BiTemporalTable	§9.1
BiTemporal Join	✅	BiTemporalJoin	§9.2
TemporalQueryEngine	✅	TemporalQueryEngine	§9.3
SnapshotManager (PITR)	✅	SnapshotManager	§9.4
Time-Travel Queries (AS OF TIMESTAMP)	✅	TimeTravelQuery	§9.5
System-Versioned Tables	✅	SystemVersionedTable	§9.6
Temporal Index Acceleration	✅	—	§9.7

Module: src/temporal/, src/timeseries/, include/temporal/, include/timeseries/
Compendium: chapter_09_timeseries.md

---

Storage & Backup (§20)

Feature	Status	Class	Documentation
RocksDB Storage Backend	✅	RocksDBStorage	§20.1
MVCC / WAL	✅	MVCCEngine, WALWriter	§20.2
BackupManager (PITR)	✅	BackupManager	§20.3
RAID-equivalent Redundancy	✅	RaidManager	§20.4
Tiered Storage (SSD→HDD→Cold)	✅	TieredStorageManager	§20.5
NVMe-aware I/O	✅	NVMeDriver	§20.6
Erasure Coding (Reed-Solomon)	✅	ErasureCoder	§20.7
Object Storage Backend (S3/GCS/Azure)	✅	ObjectStorageBackend	§20.8
2PC Distributed Storage Commit	✅	DistributedStorageCommit	§20.9

Module: src/storage/, include/storage/
Compendium: chapter_08_storage_layer.md, chapter_20_backup.md

---

LLM Integration & Inference (§17)

Feature	Status	Class	Documentation
llama.cpp Wrapper	✅	LlamaWrapper	§17.1
GGUF Model Loader	✅	GGUFLoader	§17.2
LoRA Adapter Support	✅	LoRAAdapter	§17.3
QLoRA (Quantized LoRA)	✅	QLoRATrainer	§17.4
Multi-GPU Training	✅	MultiGPUTrainer	§17.5
Paged KV-Cache	✅	PagedKVCache	§17.6
Continuous Batching	✅	ContinuousBatcher	§17.7
Speculative Decoding	✅	SpeculativeDecoder	§17.8
OpenAI-Compatible API	✅	OpenAICompatLayer	§17.9
LoRA Router	✅	LoRARouter	§17.10
Adapter Registry	✅	AdapterRegistry	§17.11
Model Router	✅	ModelRouter	§17.12
Prefix Caching	✅	PrefixCache	§17.13
Grammar-Constrained Generation	✅	GrammarConstrainedSampler	§17.14
RoPE Scaling	✅	RoPEScaling	§17.15
Vision / Multi-Modal	✅	VisionEncoder	§17.16
GPU Backend Auto-Detection	✅	AiHardwareDispatcher	§17.17
LLM Benchmarking Framework	✅	LLMBenchmark	§17.18
Feedback API	✅	FeedbackCollector	§17.19
LLM-as-Judge	✅	LLMJudge	§17.20

Module: src/llm/, src/llama_cpp/, include/llm/
Compendium: chapter_17_llm_integration.md

---

RAG — Retrieval-Augmented Generation (§17 continued)

Feature	Status	Class	Documentation
Basic RAG Pipeline	✅	RAGPipeline	§17.21
Agentic RAG (multi-step)	✅	AgenticRAG	§17.22
Multi-Modal RAG	✅	MultiModalRAG	§17.23
RAG Context Assembler	✅	RAGContextAssembler	§17.24
Distributed RAG Evaluator	✅	DistributedRAGEvaluator	§17.25
Hybrid Retrieval (BM25+Vector)	✅	HybridRetriever	§17.26
Knowledge Graph RAG	✅	KGRAGPipeline	§17.27
FLARE Retrieval-Callback Bridge	✅	KnowledgeGapDetector::setRetrievalCallback	§17.30
Vector Search in RAG (HTTP)	✅	LLMApiHandler::setVectorIndex	§17.30

Module: src/rag/, include/rag/, src/server/, include/server/
Compendium: chapter_17_llm_integration.md
Referenz: flare_retrieval_callback_bridge.md

---

Prompt Engineering (§17 continued)

Feature	Status	Class	Documentation
PromptManager (RocksDB, lock-free)	✅	PromptManager	§17.28
FeedbackCollector (10 types)	✅	FeedbackCollector	§17.29
SelfImprovementOrchestrator	✅	SelfImprovementOrchestrator	§17.30
TreeOfThoughtsBuilder	✅	TreeOfThoughtsBuilder	§17.31
ProTeGi Optimizer	✅	ProTeGiOptimizer	§17.32
DSPy Module	✅	DSPyModule	§17.33
ContextWindowManager	✅	ContextWindowManager	§17.34

Module: src/prompt_engineering/, include/prompt_engineering/
Compendium: chapter_17_llm_integration.md

---

Training & ML Pipeline (§16 ML)

Feature	Status	Class	Documentation
TrainingPipeline	✅	TrainingPipeline	§16-ML.1
IncrementalLoRATrainer	✅	IncrementalLoRATrainer	§16-ML.2
AutoLabeler	✅	AutoLabeler	§16-ML.3
LoRACheckpointManager	✅	LoRACheckpointManager	§16-ML.4
ProvenanceTracker	✅	ProvenanceTracker	§16-ML.5
Model Serving (ONNX)	✅	ModelServingEngine	§16-ML.6

Module: src/training/, include/training/
Compendium: chapter_16_ml.md

---

Change Data Capture (§11)

Feature	Status	Class	Documentation
Changefeed (row-level CDC)	✅	Changefeed	§11.1
CDC Materialized Views	✅	CdcMaterializedView	§11.2
CDC Admin API	✅	CdcAdmin	§11.3
Consumer Group	✅	ConsumerGroup	§11.4
Kafka Integration	✅	KafkaChangefeed	§11.5
Avro / Protobuf Encoding	✅	AvroEncoder, ProtoEncoder	§11.6

Module: src/cdc/, include/cdc/
Compendium: chapter_11_realtime.md

---

Update Propagation & Cluster Operations (§25)

Feature	Status	Class	Documentation
ClusterUpdateManager	✅	ClusterUpdateManager	§25.1
CanaryRollout	✅	CanaryRollout	§25.2
HotReloadEngine	✅	HotReloadEngine	§25.3
InPlaceSchemaMigrator	✅	InPlaceSchemaMigrator	§25.4
BlueGreenDeployment	✅	BlueGreenDeployment	§25.5
SIGHUP Config Hot-Reload	✅	ConfigHotReload	§25.6

Module: src/updates/, src/config/, include/updates/, include/config/
Compendium: chapter_25_devops_infrastructure.md

---

Horizontal Scaling & Sharding (§16)

Feature	Status	Class	Documentation
AdaptiveShardRouter	✅	AdaptiveShardRouter	§16.1
Consistent Hash Sharding	✅	ConsistentHashSharding	§16.2
Range Sharding	✅	RangeSharding	§16.3
Cross-Shard Transactions	✅	CrossShardTransaction	§16.4
Epoch Fencing	✅	EpochFencing	§16.5
Data Migration / Rebalancing	✅	DataMigrator	§16.6
GPU Erasure Coder (OpenCL)	✅	GpuErasureCoderOpenCL	§16.7

Module: src/sharding/, include/sharding/
Compendium: chapter_16_sharding.md

---

Query Optimization (§34)

Feature	Status	Class	Documentation
Adaptive Query Optimizer	✅	AdaptiveOptimizer	§34.1
Adaptive Join (Hash/NL/Merge/Grace)	✅	AdaptiveJoin	§34.2
Approximate Aggregator (HLL+TDigest)	✅	ApproximateAggregator	§34.3
Materialized Views (Incremental)	✅	IncrementalMaterializedView	§34.4
Parallel Query Execution	✅	ParallelExecutor	§34.5
Partition Pruning	✅	PartitionPruner	§34.6
Cost Model (System R)	✅	CostModel	§34.7
Query Plan Cache	✅	QueryPlanCache	§34.8
Early Filtering	✅	EarlyFilter	§34.9

Module: src/query/, include/query/
Compendium: chapter_34_query_optimization.md

---

Index Management (§34 continued)

Feature	Status	Class	Documentation
IndexManager	✅	IndexManager	§34.10
Inverted Index (BM25)	✅	InvertedIndex	§34.11
HNSW Vector Index	✅	HNSWIndex	§34.12
Distributed Vector Index	✅	DistributedVectorIndex	§34.13
Index Suggestion Engine	✅	IndexSuggestionEngine	§34.14
B-Tree / B+Tree Index	✅	BTreeIndex	§34.15
Spatial Index (R-Tree)	✅	RTreeIndex	§34.16

Module: src/index/, include/index/
Compendium: chapter_34_query_optimization.md

---

Full-Text & Hybrid Search (§13)

Feature	Status	Class	Documentation
HybridSearch (BM25+Vector, RRF k=60)	✅	HybridSearch	§13.1
DistributedHybridSearch (mTLS)	✅	DistributedHybridSearch	§13.2
SearchHighlighter	✅	SearchHighlighter	§13.3
SearchResultStream (cursor)	✅	SearchResultStream	§13.4
ConversationalSearch	✅	ConversationalSearch	§13.5
FederatedSearch	✅	FederatedSearch	§13.6
FacetedSearch	✅	FacetedSearch	§13.7
LearningToRank	✅	LearningToRank	§13.8
AutocompleteEngine	✅	AutocompleteEngine	§13.9
MultiModalSearch	✅	MultiModalSearch	§13.10

Module: src/search/, include/search/
Compendium: chapter_13_fulltext.md

---

Analytics (§15 / §29)

Feature	Status	Class	Documentation
OLAP Engine (CUBE/ROLLUP/GROUPING SETS)	✅	OLAPEngine	§15.1
GPU OLAP Path	✅	GPUOLAPEngine	§15.2
Window Functions	✅	WindowFunctionExecutor	§15.3
CEP Engine (NFA, EPL)	✅	CEPEngine	§15.4
Anomaly Detection (Isolation Forest)	✅	AnomalyDetector	§15.5
Model Serving (ONNX)	✅	ModelServingEngine	§15.6
Incremental Materialized Views	✅	IncrementalMatView	§15.7
Process Mining (BPMN/EPK)	✅	ProcessAnalyzer	§29.1
ExporterFactory (Arrow/Parquet)	✅	ExporterFactory	§15.8
JoinExporter (PII-redacting)	✅	JoinExporter	§15.9

Module: src/analytics/, include/analytics/
Compendium: chapter_15_analytics.md, chapter_29_analytics_process_mining.md

---

Security & Encryption (§36)

Feature	Status	Class	Documentation
Row-Level Security (PERMISSIVE + RESTRICTIVE)	✅	RLSManager	§36.1
Zero Trust Policy Enforcer	✅	ZeroTrustPolicyEnforcer	§36.2
Field Encryption (AES-256-GCM)	✅	FieldEncryption	§36.3
DEK/KEK/MasterKey Hierarchy	✅	KeyManager	§36.4
HSM / Vault Integration	✅	VaultClient	§36.5
AQL Injection Detection	✅	AQLInjectionDetector	§36.6
Malware Scanner	✅	MalwareScanner	§36.7
eIDAS Timestamping	✅	EiDASTimestamp	§36.8
OCSP / CRL Revocation Check	✅	PluginSecurityVerifier	§36.9
ArrowUserRegistrationPlugin	✅	ArrowUserRegistrationPlugin	§36.10

Module: src/security/, include/security/
Compendium: chapter_36_security_hardening.md

---

Data Governance & Compliance (§40)

Feature	Status	Class	Documentation
PolicyEngine (OPA-fallback)	✅	PolicyEngine	§40.1
DataMasker (REDACT/HASH/TRUNCATE/TOKENIZE/ENCRYPT)	✅	DataMasker	§40.2
DataLineageTracker (append-only)	✅	DataLineageTracker	§40.3
CrossTenantPolicyInheritance	✅	CrossTenantPolicy	§40.4
ComplianceReporter (GDPR/HIPAA/CCPA/PCI/SOC2/ISO27001)	✅	ComplianceReporter	§40.5
PolicyManager hot-reload	✅	PolicyManager	§40.6

Module: src/governance/, include/governance/
Compendium: chapter_40_data_governance_compliance.md

---

Transactions & Concurrency (MVCC / HLC)

Feature	Status	Class	Documentation
BranchManager (Git-style branches)	✅	BranchManager	mvcc§1
SnapshotManager (named tags, PITR)	✅	SnapshotManager	mvcc§2
SAGA Distributed Orchestration	✅	SAGAOrchestrator	mvcc§3
DeadlockPredictor (probability)	✅	DeadlockPredictor	mvcc§4
DistributedTransactionManager (2PC)	✅	DistributedTransactionManager	mvcc§5
GlobalTransactionManager (TrueTime)	✅	GlobalTransactionManager	mvcc§6
Serializable Snapshot Isolation	✅	IsolationLevel::SerializableSnapshot	mvcc§7
Savepoints	✅	Savepoint	mvcc§8
Optimistic Concurrency Control	✅	OCCValidator	mvcc§9
Bulk Transaction API	✅	BulkTransactionAPI	mvcc§10

Module: src/transaction/, include/transaction/
Compendium: chapter_mvcc_hlc.md

---

High Availability & Replication (§18)

Feature	Status	Class	Documentation
WALArchivalManager (Zstd+AES-256-GCM)	✅	WALArchivalManager	§18.1
IArchivalBackend (S3/GCS/Azure)	✅	IArchivalBackend	§18.2
Storage-Tier Lifecycle	✅	TierLifecycleManager	§18.3
LogicalReplicationManager	✅	LogicalReplicationManager	§18.4
FaultInjector (ChaosFramework)	✅	FaultInjector	§18.5
ChaosScheduler	✅	ChaosScheduler	§18.6
DisasterRecoveryManager (7-step)	✅	DisasterRecoveryManager	§18.7
Auto Failover Manager	✅	AutoFailoverManager	§18.8

Module: src/replication/, src/chaos/, src/failover/, include/replication/, include/chaos/, include/failover/
Compendium: chapter_18_ha.md

---

Observability & SRE (§38)

Feature	Status	Class	Documentation
MetricsCollector (40+ Prometheus metrics)	✅	MetricsCollector	§38.1
QueryProfiler (per-phase timing)	✅	QueryProfiler	§38.2
StorageProfiler (write/read amplification)	✅	StorageProfiler	§38.3
PerformanceAnalyzer (INFO/WARNING/CRITICAL)	✅	PerformanceAnalyzer	§38.4
AlertingEngine (PagerDuty/Slack)	✅	AlertingEngine	§38.5
Distributed Tracing (OpenTelemetry)	✅	—	§38.6
Kubernetes Health Probes	✅	—	§38.7
Grafana Dashboards (3 dashboards)	✅	—	§38.8

Module: src/observability/, include/observability/
Compendium: chapter_38_observability_sre.md

---

Importers (§26)

Feature	Status	Class	Documentation
PostgresImporter (FK-preservation)	✅	PostgresImporter	§26.1
MongoImporter	✅	MongoImporter	§26.2
MySQLImporter / MariaDB	✅	MySQLImporter	§26.3
OracleImporter	✅	OracleImporter	§26.4
SqliteImporter	✅	SqliteImporter	§26.5
FlatFileImporter (CSV/TSV/Parquet)	✅	FlatFileImporter	§26.6
KafkaImporter	✅	KafkaImporter	§26.7
S3Importer	✅	S3Importer	§26.8
SchemaInferenceEngine	✅	SchemaInferenceEngine	§26.9
AuditedImporter	✅	AuditedImporter	§26.10
IImporterPlugin API	✅	IImporterPlugin	§26.11

Module: src/importers/, include/importers/
Compendium: chapter_26_migration_legacy.md

---

Ingestion (§11 continued)

Feature	Status	Class	Documentation
IngestionBuilder (fluent API)	✅	IngestionBuilder	§11.7
HuggingFace Source	✅	—	§11.8
Filesystem / API / Kafka / S3 Sources	✅	—	§11.9
CDC Source Connector	✅	—	§11.10
Schema Validation + Quarantine	✅	—	§11.11
Rate Limiting + Dry-Run	✅	—	§11.12
Incremental Checkpoint	✅	—	§11.13
Plugin API	✅	IIngestionPlugin	§11.14

Module: src/ingestion/, include/ingestion/
Compendium: chapter_11_realtime.md

---

Exporters (§16 ML continued)

Feature	Status	Class	Documentation
JsonlExporter (Alpaca/ShareGPT/ChatML)	✅	JsonlExporter	§16.1
ParquetExporter	✅	ParquetExporter	§16.2
ArrowIpcExporter (zero-copy)	✅	ArrowIpcExporter	§16.3
HuggingFaceExporter (Hub upload)	✅	HuggingFaceExporter	§16.4
StreamingExporter (cursor+progress)	✅	StreamingExporter	§16.5
IncrementalExporter (delta+watermark)	✅	IncrementalExporter	§16.6
AES-256-GCM Encryption + PII Redaction	✅	—	§16.7

Module: src/exporters/, include/exporters/
Compendium: chapter_16_ml.md

---

Caching (§8 Storage)

Feature	Status	Class	Documentation
AdaptiveQueryCache (L1/L2/L3)	✅	AdaptiveQueryCache	§8.1
BoundedLRU + zstd/lz4	✅	—	§8.2
SemanticCache (cosine similarity)	✅	SemanticCache	§8.3
GDPR invalidatePII()	✅	—	§8.4
Tenant Isolation + Quotas	✅	—	§8.5
Circuit Breaker (L3)	✅	—	§8.6
Warmup / Snapshot Export	✅	—	§8.7
PredictivePrefetcher (Markov-chain)	✅	PredictivePrefetcher	§8.8

Module: src/cache/, include/cache/
Compendium: chapter_08_storage_layer.md

---

Network & Protocols (§31)

Feature	Status	Class	Documentation
WireProtocolServer (port 8766, mTLS)	✅	WireProtocolServer	§31.1
Wire Protocol V2 (RFC 7540 §6.3)	✅	WireProtocolV2	§31.2
QUIC Transport (port 8770, 0-RTT)	✅	QuicTransport	§31.3
UDP Fast-Path (port 8769)	✅	UDPServer	§31.4
gRPC (port 8771)	✅	GrpcServer	§31.5
RaftLoadBalancer	✅	RaftLoadBalancer	§31.6
UDP Bandwidth Management / QoS	✅	BandwidthManager	§31.7
SCRAM-SHA-256	✅	SCRAMAuthenticator	§31.8
IPv6 Dual-Stack	✅	—	§31.9

Module: src/network/, include/network/
Compendium: chapter_31_api_protocols.md

---

Content Management (§7)

Feature	Status	Class	Documentation
ingestRawBlob (10-stage pipeline)	✅	ContentManager	§7.1
PdfProcessor	✅	PdfProcessor	§7.2
OfficeProcessor	✅	OfficeProcessor	§7.3
HtmlProcessor / MarkdownProcessor	✅	HtmlProcessor	§7.4
OcrProcessor	✅	OcrProcessor	§7.5
DeduplicationChecker (pHash+MinHash)	✅	DeduplicationChecker	§7.6
LanguageDetector	✅	LanguageDetector	§7.7
ContentManagerLlm	✅	ContentManagerLlm	§7.8

Module: src/content/, include/content/
Compendium: chapter_07_document.md

---

Scheduler & Retention (§8 Scheduler)

Feature	Status	Class	Documentation
TaskScheduler (Cron/CDC/DAG/Retry)	✅	TaskScheduler	§8.10
HybridRetentionManager (3-stage)	✅	HybridRetentionManager	§8.11
DistributedTaskCoordinator	✅	DistributedTaskCoordinator	§8.12
Dynamic Worker Scaling	✅	—	§8.13

Module: src/scheduler/, include/scheduler/
Compendium: chapter_08_storage_layer.md

---

Graph Database (§6)

Feature	Status	Class	Documentation
GraphQueryOptimizer (BFS/DFS/A*/Dijkstra)	✅	GraphQueryOptimizer	§6.1
EMA Adaptive Cost Model	✅	—	§6.2
Parallel Frontier Expansion	✅	—	§6.3
Constrained Path Finding	✅	—	§6.4
DistributedGraphManager	✅	DistributedGraphManager	§6.5
EXPLAIN Endpoint	✅	—	§6.6
Streaming Graph Results	✅	—	§6.7
ScheduledGraphEdgeRefreshEngine	✅	ScheduledGraphEdgeRefreshEngine	§6.11

Module: src/graph/, include/graph/
Compendium: chapter_06_graph.md

---

Voice Assistant (§17 Voice)

Feature	Status	Class	Documentation
VoiceAssistant (STT+LLM+TTS)	🟢	VoiceAssistant	§17.35
Wake-Word Detection	🟢	—	§17.36
Meeting Protocol Generation	🟢	—	§17.37
VoiceBiometricAuthenticator	🟢	VoiceBiometricAuthenticator	§17.38
TelephonyBridge (SIP/WebRTC)	🟡	TelephonyBridge	§17.39
Browser WebSocket Streaming	🟢	—	§17.40
EmotionAnalyzer	🟡	EmotionAnalyzer	§17.41
VoiceMacro	🟢	VoiceMacro	§17.42

Module: src/voice/, src/whisper/, include/voice/
Compendium: chapter_17_llm_integration.md

---

Utilities (§utils)

Feature	Status	Class	Documentation
UUID v7 (RFC 9562)	✅	generate_uuid_v7()	utils§1
LZ4 Codec (safe+legacy API)	✅	LZ4Codec	utils§2
Streaming ZSTD Compressor	✅	ZstdStreamCompressor	utils§3
Streaming ZSTD Decompressor	✅	ZstdStreamDecompressor	utils§4

Module: src/utils/, include/utils/

---

In Development / Planned (v1.9.0 — v2.0.0)

Feature	Target	Module	Issue / PR
Chimera streaming result sets	v1.9.0	chimera	#4478
Chimera prepared statements	v1.9.0	chimera	#4478
ISO 27001 + HIPAA rule evaluators	v1.9.0	governance	#4484
Auth focused test targets	v1.9.1	auth	#4474
CDC replay/filter/batch-commit	v2.0.0	cdc	#4477
Query v2.0.0 port	v2.0.0	query	#4569
Storage v2.0.0 port	v2.0.0	storage	#4570
GPU traversal kernels (CUDA)	v2.0.0	graph	roadmap
Vulkan compute shaders (distance)	v2.0.0	acceleration	roadmap
Multi-GPU distributed training coord.	v2.0.0	training	roadmap
Token counting / context-window budget	v2.0.0	prompt_engineering	roadmap
OAuth 2.0 token refresh in connectors	v1.9.x	ingestion	#2408

---

Vollständige Modul-Übersicht (v1.8.0)

Modul	Status	Compendium-Kapitel
acceleration	🟢 MVP Complete	§1.5
analytics	✅ Production-Ready	§15, §29
api	✅ Production-Ready	§32
aql	✅ Production-Ready	§28, §32
auth	✅ Production-Ready	§21
base	✅ Production-Ready	§2
cache	✅ Production-Ready	§8
cdc	✅ Production-Ready	§11
chaos	✅ Production-Ready	§18
chimera	🔴 Alpha	§37
config	✅ Production-Ready	§25
content	✅ Production-Ready	§7
core	✅ Production-Ready	§2
ethics_ai	✅ Production-Ready	§24
exporters	✅ Production-Ready	§16-ML
failover	✅ Production-Ready	§18
geo	✅ Production-Ready	§14
governance	✅ Production-Ready	§40
gpu	✅ Production-Ready	§12
graph	✅ Production-Ready	§6
importers	✅ Production-Ready	§26
index	✅ Production-Ready	§34
ingestion	✅ Production-Ready	§11
llm	✅ Production-Ready	§17
maintenance	✅ Production-Ready	§30
metadata	✅ Production-Ready	§35
network	✅ Production-Ready	§31
observability	✅ Production-Ready	§38
performance	✅ Production-Ready	§20, §39
plugins	✅ Production-Ready	§37
process	✅ Production-Ready	§29
prompt_engineering	✅ Production-Ready	§17
query	✅ Production-Ready	§34
rag	✅ Production-Ready	§17
replication	✅ Production-Ready	§18
scheduler	✅ Production-Ready	§8
search	✅ Production-Ready	§13
security	✅ Production-Ready	§36
server	✅ Production-Ready	§31, §32
sharding	✅ Production-Ready	§16
storage	✅ Production-Ready	§8, §20
temporal	✅ Production-Ready	§9
themis	✅ Production-Ready	§2
timeseries	✅ Production-Ready	§9
training	✅ Production-Ready	§16-ML
transaction	✅ Production-Ready	mvcc
updates	✅ Production-Ready	§25
utils	✅ Production-Ready	—
voice	🟢 MVP Complete	§17

Gesamt: 46 Module
Production-Ready: 43 (93%)
MVP Complete: 2 (4%)
Alpha: 1 (2%)

---

Version History

Version	Datum	Highlights
v2.0.0	geplant	GPU kernel API stabilisierung, API /v1/ als stable surface, CDC replay/filter
v1.9.x	geplant	Chimera streaming, Governance ISO 27001/HIPAA, Auth test targets
v1.8.0	Apr 2026	Wire Protocol V2, SSI, SAGA Engine, Versioned API Routing, PredictivePrefetcher, GpuErasureCoder, MySQL Importer, CEP deadlock fix, PolicyManager hot-reload
v1.7.0	Feb 2026	Core module migration to src/themis/, module loader refactor
v1.6.0	Jan 2026	WALArchival (Zstd+AES-256-GCM), LogicalReplication, UUID v7, LZ4/ZSTD codecs
v1.5.0	Dez 2025	Comprehensive multi-model: HA, Voice, advanced LLM, PostgreSQL Wire Protocol
v1.4.0-alpha	Jan 2026	LLM integration, Performance optimierungen, HA, Monitoring (ursprüngliche Alpha)

---

Zweck dieses Appendix

Dieser Appendix bietet eine vollständige, strukturierte Übersicht aller ThemisDB-Features mit aktuellem Implementierungsstatus, Dokumentationsverweisen und Roadmap-Planung. Dies dient als zentraler Referenzpunkt für:

• Entwickler: Welche Features sind verfügbar? Wo ist der Code?
• Architekten: Welche Capabilities hat ThemisDB? Was fehlt noch?
• Projektmanager: Was ist Production-Ready? Was ist geplant?
• Nutzer: Welche Funktionen kann ich nutzen?

---

Status-Definitionen

Symbol	Status	Bedeutung
✅	Production-Ready	Vollständig implementiert, getestet (>85% Coverage), dokumentiert, performance-validated
🟢	MVP Complete	Kernfunktionalität vorhanden, stabil, aber ggf. noch Optimierungen ausstehend
🟡	In Development	Aktiv in Entwicklung, teilweise funktional, noch nicht production-grade
🟠	Design Phase	Spezifikation vorhanden, Implementierung geplant
🔵	Planned	Auf Roadmap, noch keine Spezifikation
⚪	Backlog	Idee/Anforderung dokumentiert, keine aktive Planung
🆕	Alpha	Neu in v1.4.0-alpha, noch nicht production-ready
❌	Not Planned	Bewusst nicht im Scope

---

Zusammenfassung

Aktuelle Reife (Stand Feb 2026):

• Core Database: ✅ Production-Ready (85%+ Test Coverage)
• Multi-Model Support: ✅ Production-Ready (Relational, Graph, Vector, Document, Time-Series)
• Security & Compliance: ✅ Production-Ready (DSGVO, BSI C5, Audit Logging)
• Horizontal Scaling: ✅ Production-Ready (2/4/8-Node Clusters, 91% Efficiency)
• High Availability: 🟢 MVP Complete (Hot Spare, WAL Replication, promoted from Alpha)
• Voice Assistant: 🆕 Beta (Whisper STT, Piper TTS, Meeting Protocols, Call Center Automation)
• LLM/RAG Integration: 🟢 MVP Complete (8 Features: Prefix/Response Caching, Multi-GPU, Paged Attention, LoRA, Grammar-Constrained Generation, RoPE Scaling, Vision - promoted from Alpha)
• Performance Optimizations: 🟢 MVP Complete (Flash Attention, Speculative Decoding, Continuous Batching)
• PostgreSQL Wire Protocol: 🟡 In Development (COPY, LISTEN/NOTIFY, Binary Format, Pipeline Mode - see src/server/FUTURE_ENHANCEMENTS.md)
• Enhanced Monitoring: ✅ Production-Ready (30+ Prometheus-Metriken für LLM, HA, Performance)

---

v1.5.0-dev Feature-Highlights

LLM-Integration (Kapitel 17)

Feature	Status	Maturity	Documentation	Performance
Prefix Caching	🟢 MVP Complete	Stable	Chapter 17.12.1	75% cost savings
Response Caching	🟢 MVP Complete	Stable	Chapter 17.12.2	60-80% savings
Multi-GPU Support	🟢 MVP Complete	Stable	Chapter 17.12.3	4-8x throughput
Paged Attention	🟢 MVP Complete	Stable	Chapter 17.12.4	80% memory reduction
LoRA Support	🟢 MVP Complete	Stable	Chapter 17.12.5	99% less memory
Grammar-Constrained Generation	🟢 MVP Complete	Stable	Chapter 17.12.6	95-99% valid outputs
RoPE Scaling	🟢 MVP Complete	Stable	Chapter 17.12.7	4K→32K context
Vision Support	🟢 MVP Complete	Stable	Chapter 17.12.8	Multimodal AI

Roadmap:

• Production-Ready (v1.6): Extended testing, performance tuning, production certification

Enterprise Features (Kapitel 10)

Feature	Status	Maturity	Documentation	Use Case
Voice Assistant	🟡 Beta	Improving	Chapter 10.7	Call Center, Meeting Protocols
Multi-Tenancy	✅ Production-Ready	Stable	Chapter 10.1	SaaS Applications
RBAC	✅ Production-Ready	Stable	Chapter 10.1	Enterprise Security
Audit Logging	✅ Production-Ready	Stable	Chapter 10.1	Compliance

Roadmap (Voice Assistant):

• Production-Ready (v1.6): Enhanced language support, improved accuracy, production certification

Performance-Optimierungen (Kapitel 21)

Feature	Status	Maturity	Documentation	Performance
Flash Attention	🟢 MVP Complete	Stable	Chapter 20.9A.1	69% throughput, 37% memory
Speculative Decoding	🟡 In Development	Experimental	Chapter 20.9A.2	2-3x speedup (target)
Continuous Batching	🟢 MVP Complete	Stable	Chapter 20.9A.3	176% throughput

Roadmap:

• Production-Ready (v1.6): Hardware compatibility testing, production deployment guides

High Availability (Kapitel 16)

Feature	Status	Maturity	Documentation	Performance
Hot Spare	🟢 MVP Complete	Stable	Chapter 16.10.1	<5s failover
WAL Replication	🟢 MVP Complete	Stable	Chapter 16.10.2	Zero data loss
Multi-SSD WAL	🆕 Alpha	Early	Chapter 16.10.2	<10% write overhead
Replication Slots	🆕 Alpha	Early	Chapter 16.10.2	Lag monitoring

Roadmap:

• Beta (Feb 2026): Extended failover testing, disaster recovery
• GA (Mar 2026): Production HA certification

Monitoring & Observability (Kapitel 19)

Feature	Status	Maturity	Documentation	Metrics Count
LLM Metrics	🆕 Alpha	Early	Chapter 19.6A.1	15+ metrics
Performance Metrics	🆕 Alpha	Early	Chapter 19.6A.2	10+ metrics
HA Metrics	🆕 Alpha	Early	Chapter 19.6A.3	8+ metrics
Grafana Dashboards	🆕 Alpha	Early	Chapter 19.6A	3 dashboards
Alerting Rules	🆕 Alpha	Early	Chapter 19.6A.4	12+ rules

Roadmap:

• Beta (Feb 2026): Dashboard refinement, extended alerting
• GA (Mar 2026): Complete observability stack

PostgreSQL Wire Protocol (Kapitel 31)

Feature	Status	Maturity	Documentation	Performance
COPY Protocol	🆕 Alpha	Early	Chapter 31.9A.1	250K rows/s
LISTEN/NOTIFY	🆕 Alpha	Early	Chapter 31.9A.1	Real-time CDC
Binary Format	🆕 Alpha	Early	Chapter 31.9A.2	80% size reduction
Pipeline Mode	🆕 Alpha	Early	Chapter 31.9A.2	17x throughput
Prepared Stmt Cache	🆕 Alpha	Early	Chapter 31.9A.2	50x speedup
Vector Type Mapping	🆕 Alpha	Early	Chapter 31.9A.3	Native pgvector
JSONB Support	🆕 Alpha	Early	Chapter 31.9A.3	Full compatibility

Roadmap:

• Beta (Feb 2026): Extended client library testing
• GA (Mar 2026): Full PostgreSQL compatibility certification

---

Vollständige Feature-Matrix

Vollständige Feature-Matrix siehe:

• Admin Tools: feature_matrix.md
• Features Overview: features_overview.md
• Sharding Status: Chapter 16 (Horizontal Scaling)
• Security Status: Chapter 10 (Section 10.2)
• Query Optimization: Chapter 15 (Section 15.4)

---

Feature-Maturity-Levels

Alpha (v1.4.0-alpha)

Definition: Neue Features in früher Entwicklung, für Feedback und Testing.

Charakteristiken:

• ✅ Kernfunktionalität implementiert
• ✅ Dokumentation vorhanden
• ⚠️ Noch nicht production-ready
• ⚠️ API kann sich ändern
• ⚠️ Performance noch nicht final optimiert

Empfohlene Nutzung:

• Development/Staging Environments
• Feature Evaluation
• Feedback und Bug Reports

Beta (geplant: Feb 2026)

Definition: Features sind stabil, werden für Production vorbereitet.

Charakteristiken:

• ✅ Vollständig implementiert
• ✅ >80% Test Coverage
• ✅ Performance-validated
• ⚠️ Noch unter intensivem Testing
• ⚠️ Minor API changes möglich

Empfohlene Nutzung:

• Pre-Production Testing
• Pilot-Deployments
• Performance Benchmarking

GA (General Availability - geplant: Mar 2026)

Definition: Production-ready, vollständig getestet und dokumentiert.

Charakteristiken:

• ✅ Production-Ready
• ✅ >85% Test Coverage
• ✅ Performance-validated
• ✅ Complete Documentation
• ✅ Stable API
• ✅ Long-term Support

Empfohlene Nutzung:

• Production Deployments
• Mission-Critical Workloads
• Enterprise Applications

---

Implementierungsstatus v1.4.0-alpha

Gesamt-Übersicht

Kategorie	Features	Alpha	Beta	GA	Geplant
LLM Integration	6	6	0	0	0
Performance	3	3	0	0	0
High Availability	4	4	0	0	0
Monitoring	5	5	0	0	0
PostgreSQL Protocol	7	7	0	0	0
Gesamt	25	25	0	0	0

Prozentsatz:

• Alpha Features: 100% (25/25)
• Production-Ready: 0% (Target: 100% by Mar 2026)

---

Implementierungsstatus v1.9.x (2026-04-12)

Die folgenden Komponenten wurden mit dem v1.9.x-Release (Commit 2026-04-12) in den Produktionscode aufgenommen:

Storage & Ingest

Feature	Modul	Status	Tests	Beschreibung
StreamingIngestManager	storage/	✅ GA	SM-01..SM-10	Ring-Buffer-Ingest ≥ 1 M Events/s, OverflowPolicy BLOCK/DROP
TSStore::putBatch	timeseries/	✅ GA	TB-01..TB-14	Zero-Copy Batch Write via std::span<const TSRow>
TsStreamCursor	timeseries/	✅ GA	SC-01..SC-08	Lazy paginierter Iterator, page_size=4 096
TemporalCompressor LZ4	temporal/	✅ GA	TC-LZ4-01..TC-LZ4-05	LZ4-Block-Kompression im TemporalCompressor
STORAGE_COMPACTION Wiring	maintenance/	✅ GA	—	StorageCompactionHandler in maintenance_orchestrator_ registriert
MVCC_CLEANUP Wiring	maintenance/	✅ GA	—	MvccCleanupHandler mit 24 h Watermark registriert

Maintenance

Feature	Modul	Status	Tests	Beschreibung
DatabaseMaintenanceOrchestrator	maintenance/	✅ GA	40+ (OrchestratorTests)	Zentraler Wartungskoordinator: CRUD, Cron, DAG, Window-Enforcement, Audit
MaintenanceScheduleStore	maintenance/	✅ GA	—	RocksDB-Persistenz für Schedules (maint_sched::<id>)
StorageCompactionHandler	maintenance/	✅ GA	—	Kapselt CompactionManager::compactAll()
MvccCleanupHandler	maintenance/	✅ GA	—	Bereinigt MVCC-Versionen mit konfigurierbarem Watermark
DAG Task Dependencies	maintenance/	✅ GA	—	depends_on-Felder; topologische Sortierung via Kahn; Zykelerkennung
Maintenance Window Enforcement	maintenance/	✅ GA	—	Tasks außerhalb window_utc_hours → SKIPPED
11 REST Endpoints	maintenance/	✅ GA	—	/api/v1/maintenance/schedules + /jobs + /status

Analytics

Feature	Modul	Status	Tests	Beschreibung
ColumnarCache	storage/	✅ GA	CC-01..CC-12	LRU-Eviction + PinGuard RAII, ≥ 10× Speedup
IStreamingJoin / HashJoin	analytics/	✅ GA	SJ-01..SJ-15	Equi-Join, Inner/LeftOuter, composite keys
IntervalJoin	analytics/	✅ GA	SJ-01..SJ-15	Zeitbasierter Event-Join, LRU-Pruning

Performance & Netzwerk

Feature	Modul	Status	Tests	Beschreibung
LockFreeHistogram<T>	performance/	✅ GA	LFH-01..LFH-12	Lock-free, ≤ 20 ns record(), Exponential+Linear
RequestCoalescer (Singleflight)	cache/	✅ GA	RC-01..RC-14	Thundering-Herd-Schutz, fn() exakt einmal
IoUringBatchedSender	network/	✅ GA	IUB-01..IUB-12	O(1) Syscalls/Runde, writev(2)-Fallback
LIRS shared_mutex Fix	performance/	✅ GA	—	TOCTOU-Race beseitigt; get() → unique_lock
RCU g_rcu_reader_count Fix	performance/	✅ GA	—	readers_active() war immer false, jetzt korrekt

Geo-Modul (v1.x)

Feature	Modul	Status	Tests	Beschreibung
ST_BUFFER	geo/	✅ GA	—	CPU-exact + Boost; GPU-Fallback mit Audit-Log
ST_UNION / ST_DIFFERENCE	geo/	✅ GA	—	CPU-exact, Boost, GPU-fallback; AQL-Funktionen
CUDA Geo-Kernel	geo/acceleration/	✅ GA	—	Haversine-Distanz, Containment-Batch (CUDA + ROCm/HIP)
S2-Zell-Indizierung	geo/	✅ GA	—	Hierarchische S2-Zellen; cellForPoint(), coveringCells()
H3-Hexagonales Grid	geo/	✅ GA	—	Uber H3; geoToH3(), kRing(), compact()
TemporalSpatialQuery	geo/	✅ GA	—	Location-at-Time T; locationAtTime(), entitiesInBboxAtTime()
RasterGrid / sampleAt()	geo/	✅ GA	—	Elevation-Sampling mit bilinearer Interpolation
generateHeatmap()	geo/	✅ GA	—	Gaussian-KDE aus Punktwolke; konfigurierbare Bandbreite + Auflösung
Spatial JOIN	geo/	✅ GA	—	Alle Paare innerhalb Distanz (CUDA-optimiert)
ROCm/HIP-Backend	geo/	✅ GA	—	AMD GPU-Unterstützung als CUDA-Alternative

AI / Acceleration

Feature	Modul	Status	Tests	Beschreibung
AiHardwareDispatcher	acceleration/	✅ GA	AHD-01..AHD-30	NPU-Prioritätskette, INT4/W4A8, ONNX-EP-Fallback
GPUVectorIndex Oversubscription	index/	✅ GA	26 (GPUMemOversubTests)	Partitions-basiertes VRAM-Paging; LRU/MRU/SEQUENTIAL Prefetch
GPUMemoryOversubscriptionManager	index/	✅ GA	26 (GPUMemOversubTests)	Unified Memory (cudaMallocManaged/hipMallocManaged); CPU-Fallback
GPUUnifiedMemoryAllocator	index/	✅ GA	—	Transparente CUDA/HIP Unified Memory + CPU-Heap-Fallback
CUDAVectorBackend	acceleration/	✅ GA	—	HNSW on GPU; Batch-KNN; CUDA Graph Capture; buildHnswAnnIndex
MultiGPUVectorBackend	acceleration/	✅ GA	—	NCCL/RCCL Fan-out KNN; NVLink P2P; CPU-Fallback; bis zu N GPUs
GeoAccelerationBridge	acceleration/	✅ GA	—	CUDA/Vulkan Haversine + Point-in-Polygon; Backend-Dispatch
Vulkan Compute (non-NVIDIA)	acceleration/	✅ GA	—	SPIR-V Shaders: L2/Cosine/InnProd/TopK/Batch+Geo-Kernels
ROCm/HIP Backend (AMD)	acceleration/	✅ GA	—	ann_kernels.hip + geo_kernels.hip; non-HIP CPU-Fallback

Utils

Feature	Modul	Status	Tests	Beschreibung
UUID v7 (generate_uuid_v7)	utils/	✅ GA	UV7-01..UV7-20	RFC 9562, 48-bit Timestamp, 18-bit Seq, MT19937-64
Streaming ZSTD (zstd_compress_stream)	utils/	✅ GA	ZS-01..ZS-10	ZSTD_CStream/DStream, max_output_bytes DoS-Guard

Content-Modul (v1.x)

Feature	Modul	Status	Tests	Beschreibung
ContentManager::ingestRawBlob()	content/	✅ GA	—	Auto-MIME, 10-stufige Pipeline, Archive-Rekursion
PdfProcessor	content/	✅ GA	—	Text-Extraktion via poppler-cpp
OfficeProcessor (OOXML/ODF)	content/	✅ GA	—	DOCX/XLSX/PPTX/ODF via libzip+pugixml
OfficeProcessor (Legacy)	content/	✅ GA	—	DOC/XLS/PPT via LibreOffice-Headless-Fallback
HtmlProcessor	content/	✅ GA	—	Boilerplate-Entfernung
MarkdownProcessor	content/	✅ GA	—	Frontmatter-Parsing
SttProcessor / TtsProcessor	content/	✅ GA	—	Whisper-Transkription + TTS
OcrProcessor	content/	✅ GA	—	Tesseract OCR für Bild-PDFs
DeduplicationChecker	content/	✅ GA	—	pHash (Bilder) + MinHash+LSH (Text); Jaccard-Schwellenwert
LanguageDetector	content/	✅ GA	—	Multi-Sprach-Erkennung mit Konfidenz
ContentManagerLlm	content/	✅ GA	—	Summary, Topics, Sentiment, Category via LLM

Scheduler-Modul (v1.5)

Feature	Modul	Status	Tests	Beschreibung
TaskScheduler	scheduler/	✅ GA	—	Cron/Interval/CDC/Manual/Webhook Trigger; 5/6-Feld, Timezone
DAG-Workflow-Engine	scheduler/	✅ GA	—	Topologische Ausführung, paralleler Fan-Out, branch_condition
Retry-Policies	scheduler/	✅ GA	—	FIXED/EXPONENTIAL/LINEAR/JITTER/FIBONACCI
TaskScheduler Dynamic Scaling	scheduler/	✅ GA	—	Auto-Concurrency via Queue-Tiefe + enable_dynamic_scaling
TaskAuditManager	scheduler/	✅ GA	—	Durchsuchbares Audit-Log; GDPR-Modus
TaskAnomalyDetector	scheduler/	✅ GA	—	Anomalie-Erkennung über Task-Ausführungs-Metriken
Prometheus-Export (exportMetrics())	scheduler/	✅ GA	—	Metriken-Text für Scraping
HybridRetentionManager	scheduler/	✅ GA	—	3-Stufen: Gorilla (0–7d), Varianz-Downsampling (7–365d), Tages-Aggregate (>365d)
DistributedTaskCoordinator	scheduler/	✅ GA	—	Leader-Election; One-Runner-Per-Cluster
ExternalSchedulerAdapter	scheduler/	✅ GA	—	Kubernetes CronJob + Apache Airflow Adapter

Analytics-Modul (v1.9)

Feature	Modul	Status	Tests	Beschreibung
OLAPEngine GPU-Pfad	analytics/	✅ GA	—	GPU-beschleunigt (threshold=10K rows); LRU Result Cache (TTL)
OLAPEngine CUBE/ROLLUP	analytics/	✅ GA	—	CUBE (alle Dimensionskombinationen), ROLLUP (Hierarchisch), GROUPING SETS
Window Functions	analytics/	✅ GA	—	ROW_NUMBER, SUM/AVG OVER, P50/P90/P99, ROWS/RANGE BETWEEN
CEPEngine (EPL Parser)	analytics/	✅ GA	—	NFA Pattern Matching; EPL CREATE RULE; WINDOW/GROUP BY/HAVING/ACTION
CEPEngine Event Types	analytics/	✅ GA	—	CDC (INSERT/UPDATE/DELETE), Graph, Query, System events; MPMC Ring Buffer
AnomalyDetector	analytics/	✅ GA	—	Isolation Forest; konfigurierbare Kontaminationsrate; Rolling Window
ModelServingEngine	analytics/	✅ GA	—	Online + Batch ONNX Inference; versioniertes Model Loading
Inkrementelle Materialized Views	analytics/	✅ GA	—	CDC-getrieben; Delta-Update ohne Full-Recompute

Search-Modul (v2.4)

Feature	Modul	Status	Tests	Beschreibung
HybridSearch	search/	✅ GA	—	BM25+Vector RRF (k=60); konfig. Gewichte; Partial-Result Stats
DistributedHybridSearch	search/	✅ GA	—	Cross-Shard RRF; mTLS; skip_failed_shards Degraded Mode
SearchHighlighter	search/	✅ GA	—	Term Highlighting <mark>; Best-Passage Snippet Extraktion
SearchResultStream	search/	✅ GA	—	Cursor-basierte Streaming-Pagination (total_k/page_size)
ConversationalSearch	search/	✅ GA	—	Multi-Turn Context Window (last N turns); automatische Query-Reformulation
FederatedSearch	search/	✅ GA	—	Tenant-isolierte Parallel-Suche mit per-Tenant Gewichtung
FacetedSearch	search/	✅ GA	—	Facettierte Suche über kategorische + Bereichs-Felder
LearningToRank	search/	✅ GA	—	Listwise/Pairwise LTR; Feature-basiertes Re-Ranking
AutocompleteEngine	search/	✅ GA	—	Prefix-Trie Autocomplete; Fuzzy + Phonetik
MultiModalSearch	search/	✅ GA	—	Text + Bild + Audio Query Fusion

Ingestion-Modul (v1.5)

Feature	Modul	Status	Tests	Beschreibung
IngestionManager	ingestion/	✅ GA	—	Parallele Multi-Source-Orchestrierung; Thread-Pool; Prometheus-Metriken
IngestionBuilder	ingestion/	✅ GA	—	Fluent API (withKafkaSource/withApiSource/withHuggingFaceSource/...)
GenericApiConnector	ingestion/	✅ GA	—	Cursor/Offset-Pagination; OAuth 2.0; libcurl; Exponential Back-off
HuggingFaceConnector	ingestion/	✅ GA	—	Dataset-Split-Ingest; API-Token-Auth
ObjectStorageConnector	ingestion/	✅ GA	—	S3/GCS/Azure Blob; Prefix-Filter
KafkaConnector	ingestion/	✅ GA	—	Consumer-Group; Auto-Commit; Back-pressure
WebCrawlerConnector	ingestion/	✅ GA	—	Sitemap + Recursive Crawl; politeness_delay
CdcConnector (PostgreSQL)	ingestion/	✅ GA	—	Logical Replication Slot; WAL-basiertes CDC
DatabaseConnector (JDBC/ODBC)	ingestion/	✅ GA	—	Batch-Select + Cursor-basierte Pagination
Quarantine + Retry	ingestion/	✅ GA	—	Exponential Back-off; Dead-Letter nach max_retries
Incremental Checkpoint	ingestion/	✅ GA	—	Fortschritt persistiert; Re-Start ab letztem Checkpoint
Plugin API	ingestion/	✅ GA	—	ISourceConnector-Interface für Third-Party Konnektoren

Importers-Modul (v2.1)

Feature	Modul	Status	Tests	Beschreibung
PostgresImporter v2.1	importers/	✅ GA	—	FK-Preservation (ON DELETE/UPDATE, DEFERRABLE); CHECK/EXCLUDE/GENERATED Constraints
MongoImporter	importers/	✅ GA	—	BSON→JSON; _id-Preservation
MySQLImporter	importers/	✅ GA	—	Batch-Import; Charset-Mapping
OracleImporter	importers/	✅ GA	—	DDL + Data Export
SqliteImporter	importers/	✅ GA	—	Lightweight; kein Server nötig
FlatFileImporter	importers/	✅ GA	—	CSV/TSV/Parquet; Auto-Type-Detection; Parquet-Kompression
KafkaImporter	importers/	✅ GA	—	Consumer-Group; Offset-Tracking
S3Importer	importers/	✅ GA	—	Prefix-Filter; Multipart-Download
IImporter::importDataStreaming()	importers/	✅ GA	—	Low-Memory Streaming; per-Row Callback; frühzeitiger Abbruch
SchemaInferenceEngine	importers/	✅ GA	—	Implicit FK Discovery; Semantic Type Detection; Cardinality Estimation
AuditedImporter	importers/	✅ GA	—	Audit-Trail für jede importierte Zeile + Schema-Änderung
IImporterPlugin API	importers/	✅ GA	—	Plugin-Interface für Third-Party Importer

---

Implementierungsstatus — Plugins & Sharding

Plugins (v2.x)

Feature	Modul	Status	Tests	Beschreibung
WhisperPlugin v2.0	whisper/	✅ GA	30 (WhisperPluginFocusedTests)	IAudioBackend, WavAudioChunkReader, Strategy Pattern, Provenienz-Stempel
WavAudioChunkReader	whisper/	✅ GA	im WhisperPlugin-Suite	RIFF/WAV ohne libsndfile; 16-bit PCM + float32
SDPlugin v2.1	stable_diffusion/	✅ GA	45 (SDPluginFocusedTests, A–O)	Text2Img, Batch, Img2Img, SDPromptSanitizer, Provenienz
SDStubGenerator	stable_diffusion/	✅ GA	im SDPlugin-Suite	CI-Stub ohne Modell-Datei
SDPromptSanitizer	stable_diffusion/	✅ GA	im SDPlugin-Suite	Keyword-Blocklist, negative_prompt-Policy (SD-NP-01)

Network-Modul (v1.8)

Feature	Modul	Status	Tests	Beschreibung
WireProtocolServer	network/	✅ GA	—	Port 8766; dedizierted I/O+Worker-Pool; mTLS; SCRAM-SHA-256
WireProtocolServer WebSocket-Upgrade	network/	✅ GA	—	HTTP-Upgrade auf Port 8766; JSON Text-Frames (THEMIS_ENABLE_WEBSOCKET)
WireProtocolServer IPv6+Dual-Stack	network/	✅ GA	—	IPv6-Socket + ipv6_dual_stack für IPv4-mapped Verbindungen
QuicTransport	network/	✅ GA	—	Port 8770; UDP+TLS 1.3; 0-RTT; kein HOL-Blocking
RaftLoadBalancer	network/	✅ GA	—	Raft In-Process-Konsensus; konsistentes Routing; Leader-Election
UDP Fast-Path	network/	✅ GA	—	Port 8769; latenzoptimierter Ingest-Pfad
gRPC Transport	network/	✅ GA	—	Port 8771; HTTP/2 Multiplexing

Security-Modul (v1.x)

Feature	Modul	Status	Tests	Beschreibung
RLSManager	security/	✅ GA	—	PostgreSQL-kompatible Row-Level Security; PERMISSIVE+RESTRICTIVE; JSON-Persistenz
ZeroTrustPolicyEnforcer	security/	✅ GA	—	Per-Request Identity + CIDR Whitelist/Blacklist; Composite Trust Score
FieldEncryption (AES-256-GCM)	security/	✅ GA	—	Field-Level Encryption; Key Rotation (90d); DEK/KEK/MasterKey Hierarchie
HSM Integration	security/	✅ GA	—	HashiCorp Vault; HSM-backed Key Storage
AQL Injection Detection	security/	✅ GA	—	Pattern-basierte Angriffserkennung in AQL-Queries
Malware Scanner (Plugin Manifests)	security/	✅ GA	—	Plugin-Manifest-Scan vor Activation
CMS/PKCS#7 + eIDAS Timestamping	security/	✅ GA	—	Digitale Signaturen; eIDAS-konformes Timestamping

Feature	Modul	Status	Tests	Beschreibung
EthicsEvaluator	ethics_ai/	✅ GA	28 (EthicsAiPluginFocusedTests)	5-Dimensions-Scoring (Quality/Consistency/Fairness/Alignment/Transparency)
PhilosophyLoader	ethics_ai/	✅ GA	7 (PhilosophyLoaderFocusedTests)	YAML-Profilladung mit Cache und Validierung
EthicalDiscourseEngine	ethics_ai/	✅ GA	11 (DiscourseEngineFocusedTests)	initializeDebate() + makeDecision()
ArgumentStore	ethics_ai/	✅ GA	18 (ArgumentStoreStandaloneFocusedTests)	BaseEntity-Persistenz + Standalone-Modus
RAGContextEngine	ethics_ai/	✅ GA	13 (RAGContextEngineFocusedTests)	7 AQL-Abfragemuster
Ethics AI Integration	ethics_ai/	✅ GA	21 (EthicsAiIntegration)	FullPipeline, ArgumentStoreRAG, RAGContextBuild

LLM Module — Produktionskomponenten (v1.0)

Feature	Modul	Status	Tests	Beschreibung
LlamaWrapper::generateVision()	llm/	✅ GA	—	Multi-Modal Inference via CLIP-VisionEncoder; LLaVA-Architektur
VisionEncoder	llm/	✅ GA	—	CLIP-GGUF-Modell; encodeImage() → float-Embedding-Vektor
LlamaWrapper RoPE Scaling	llm/	✅ GA	—	LINEAR/NTK/YARN/DYNAMIC; 4K→32K Kontext-Extension
MultiLoRAManager::loadLoRA()	llm/	✅ GA	—	vLLM-Stil; dynamic load, INT8/INT4-Quantisierung via quantizeLoRA()
MultiLoRAManager Multi-GPU	llm/	✅ GA	—	ROUND_ROBIN/DATA_PARALLEL/MODEL_PARALLEL; GPUDirect P2P
ProductionValidator	llm/testing/	✅ GA	—	72h-Stresstest, Lasttest, Performance-Regression
IntegrationTestSuite	llm/testing/	✅ GA	—	14 Szenarien (Komponenten-, Multi-Modell-, Fehler-, Performance-Tests)
LlamaWrapper Output-Validierung	llm/	✅ GA	—	UTF-8-Prüfung, min/max-Länge, Kohärenz-Score

Prompt-Engineering-Modul (v2.0)

Feature	Modul	Status	Tests	Beschreibung
PromptManager	prompt_engineering/	✅ GA	—	RocksDB-Persistenz; CRUD; TBB lock-free Hash-Map; YAML Bulk-Load
PromptManager::validateTemplate()	prompt_engineering/	✅ GA	—	Syntaxprüfung + Warnungen; in createTemplate() integriert
PromptManager::buildMultiModalPrompt()	prompt_engineering/	✅ GA	—	Multi-Modal Prompts mit ImageDescription-Blöcken
FeedbackCollector	prompt_engineering/	✅ GA	—	10 Feedback-Typen; FNV-1a Audit-Checksum; Z-Score Outlier
Git-ähnliche Versionskontrolle	prompt_engineering/	✅ GA	—	Branches, Commits, Diffs, Parent-Tracking
A/B Testing	prompt_engineering/	✅ GA	—	Welch-t-Test Signifikanztests; std::erfc-basierter Z-Score
SelfImprovementOrchestrator	prompt_engineering/	✅ GA	—	Auto-Optimierung bei Feedback-Verschlechterung; Hintergrund-Worker
TreeOfThoughtsBuilder	prompt_engineering/	✅ GA	—	Multi-Pfad-Reasoning (Beam Search, max_depth, beam_width)
ProTeGiOptimizer	prompt_engineering/	✅ GA	—	Textual-Gradient-Optimierung; population_size + iterations
DSPy-Modul (dspy_module.h)	prompt_engineering/	✅ GA	—	DSPy-kompatibler Prompt-Deklarations-Layer
ContextWindowManager	prompt_engineering/	✅ GA	—	Context-Window-Budget-Enforcement

Voice-Modul (v1.1)

Feature	Modul	Status	Tests	Beschreibung
VoiceAssistant	voice/	✅ GA	—	STT (Whisper) + LLM (LlamaWrapper) + TTS; Session-Management
Wake-Word-Detektion	voice/	✅ GA	—	"Hey Themis"; konfig. Wake-Words; Low-Latency Detection
VoiceBiometricAuthenticator	voice/	✅ GA	—	Sprecher-Enrollment + Verification; Confidence Score
TelephonyBridge (SIP/WebRTC)	voice/	✅ GA	—	Anruf-Aufzeichnung + Transkription; Sprecher-Diarisierung
Browser WebSocket Streaming	voice/	✅ GA	—	Real-Time Audio-Chunking; ws://server/voice/stream
Meeting-Protokoll-Generierung	voice/	✅ GA	—	Zusammenfassung + Action-Items + Key-Points via LLM
EmotionAnalyzer	voice/	✅ GA	—	Emotionserkennung aus Audio-Features
VoiceMacro	voice/	✅ GA	—	Konfigurierbare Sprach-Makros + Trigger-Patterns

RAG-Modul v2 (v2.x)

Feature	Modul	Status	Tests	Beschreibung
RAGJudge	rag/	✅ GA	—	5-dimensionale Evaluierung (Faith/Rel/Compl/Coh/Ethics); FAST/BALANCED/THOROUGH
HybridRetriever	rag/	✅ GA	—	BM25+Vector RRF-Fusion (k=60); linear Fallback; Factory-Helfer
DocumentSplitter	rag/	✅ GA	—	FIXED/SENTENCE/SEMANTIC/RECURSIVE; chunk_size + overlap
BatchEvaluator	rag/	✅ GA	—	Parallele Massenevaluierung mit Worker-Threads + Futures
EvaluationCache	rag/	✅ GA	—	Thread-sicherer LRU-Cache mit TTL; Invalidierungs-Triggers
CalibrationManager	rag/	✅ GA	—	Temperature Scaling, Platt Scaling, Isotonic Regression; ECE/Brier
HallucinationDashboard	rag/	✅ GA	—	Rolling-Window Halluzinationsrate + Trend (IMPROVING/STABLE/DEGRADING)
StreamingRetriever	rag/	✅ GA	—	Inkrementelles Kontext-Window-Filling
KnowledgeGraphRetriever	rag/	✅ GA	—	Entity-Linking + Graph-augmentierte Retrieval
CrossEncoderReranker	rag/	✅ GA	—	Heuristischer Scorer + ONNX-Stub
ContinuousLearningOrchestrator	rag/	✅ GA	—	Bayesian Optimierung über top_k/threshold via User-Feedback
RlaifTrainer	rag/	✅ GA	—	Constitutional AI / RLAIF Training Pipeline

Graph-Modul (v1.x)

Feature	Modul	Status	Tests	Beschreibung
GraphQueryOptimizer	graph/	✅ GA	—	Cost-Based Algo-Selektion (BFS/DFS/BiDir/A*/Dijkstra); Schema-aware
Constrained Path Finding	graph/	✅ GA	—	max_depth, required/forbidden vertices, timeout_ms
Adaptive Kostenmodell	graph/	✅ GA	—	EMA-basiertes Lernen aus Ausführungsfeedback
Parallele Traversierung	graph/	✅ GA	—	Parallele BFS-Frontier-Expansion; fan_out_threshold
DistributedGraphManager	graph/	✅ GA	—	Shard-übergreifende Graph-Queries; EXPLAIN-Endpunkt
Inkrementelle Graph-Queries	graph/	✅ GA	—	Live-Updates ohne Full-Recompute
Graph Query Result Streaming	graph/	✅ GA	—	Cursor-basierte Streaming-Pagination für große Path-Sets
ScheduledGraphEdgeRefreshEngine	graph/	✅ GA	test_scheduled_edge_refresh.cpp	Semantischer Kanten-Refresh; Kosinus/Dot/Euklid; Temporal Decay; Centrality; ACID-Batch; ANN; CEP; Changefeed; §6.11

Chaos Engineering & Failover (v1.x)

Feature	Modul	Status	Tests	Beschreibung
FaultInjector	chaos/	✅ GA	—	NODE_FAILURE/NETWORK_PARTITION/LEADER_CRASH/DISK_FAILURE/RANDOM; zeitlich begrenzt
ChaosScheduler	chaos/	✅ GA	—	Zeitgesteuerte Fault-Injections; Expiry-Handling
DisasterRecoveryManager	failover/	✅ GA	—	DR-Plan-Ausführung (7 Schritte); Step-Hooks; dry_run; Statistiken
Automatic Failover Queue	failover/	✅ GA	—	Worker-Loop; Epoch-Fencing; Catchup-Wait

Sharding / Paxos

Feature	Modul	Status	Tests	Beschreibung
Paxos WAL logAccept	sharding/	✅ GA	—	ACCEPT vor Quorum-Broadcast persistiert
Paxos WAL logCommit	sharding/	✅ GA	—	COMMIT in broadcastCommit() persistiert
recoverFromWAL()	sharding/	✅ GA	—	Snapshot + WAL-Replay; commit_index_ wird restauriert
Snapshot-Compaction	sharding/	✅ GA	—	shouldCreateSnapshot(ops) nach konfigurierbaren Operationen

Replikation (v1.6)

Feature	Modul	Status	Tests	Beschreibung
WALArchivalManager	replication/	✅ GA	—	Zstd + AES-256-GCM; S3/GCS/Azure via IArchivalBackend
Storage-Tier Lifecycle	replication/	✅ GA	—	standard → cold → glacier; konfigurierbare Alters-Schwellen
LogicalReplicationManager	replication/	✅ GA	—	Schema-aware Slots; Row-Filter, DDL-Streaming, Parallel Decoding
LogicalReplicationManager Cross-Version	replication/	✅ GA	—	Per-Change-Transformer + Cross-Version-Transforms
IArchivalBackend	replication/	✅ GA	—	Pluggable Backend-Interface für WAL-Segment-Storage

Governance-Modul (v1.x)

Feature	Modul	Status	Tests	Beschreibung
PolicyEngine::evaluate()	governance/	✅ GA	—	YAML-Policy, RBAC/ABAC, CCPA-Opt-Out, Hot-Reload
PolicyEngine::simulateDecision()	governance/	✅ GA	—	Dry-Run ohne Audit-Eintrag; matched_rule + matched_profile
PolicyEngine::checkQueryPermission()	governance/	✅ GA	—	Decision + FieldMaskingPolicy für DataMasker
PolicyEngine::checkInferencePermission()	governance/	✅ GA	—	LLM-Endpunkt Zugangskontrolle; OpenAI-style Fehlerkörper
OpaAdapter	governance/	✅ GA	—	OPA (Open Policy Agent) Integration; nativer Fallback
DataMasker::maskFields()	governance/	✅ GA	—	REDACT/HASH/TRUNCATE/TOKENIZE/ENCRYPT; letzter In-Process-Schutz
DataLineageTracker	governance/	✅ GA	—	Append-only Provenienz; Prometheus-Counter; Audit-Log
ModelGovernancePolicy	governance/	✅ GA	—	KI/ML Training Data Lineage, Bias-Auditing, Export-Check
CrossTenantPolicyInheritance	governance/	✅ GA	—	Hierarchische Policy-Vererbung über Tenants
ComplianceReporter	governance/	✅ GA	—	GDPR/HIPAA/CCPA/PCI-DSS/SOC 2/ISO 27001 Reports (JSON + PDF)

Transaction-Modul (v1.9)

Feature	Modul	Status	Tests	Beschreibung
BranchManager	transaction/	✅ GA	—	Git-ähnliche Branches + 3-Way-Merge; from_tag/from_sequence
SnapshotManager	transaction/	✅ GA	—	Named Snapshots/Tags; PITR-Basis
SAGA Distributed	transaction/	✅ GA	—	Kompensationsaktionen für Relational/Index/Graph/Vector; Distributed Orchestration
DeadlockPredictor	transaction/	✅ GA	—	Probability Scoring (0–1); Lock-Order-Empfehlung; Adaptive Timeout
DistributedTransactionManager (2PC)	transaction/	✅ GA	—	WAL-backed Coordinator; Parallel Prepare/Commit; Crash Recovery
GlobalTransactionManager (TrueTime)	transaction/	✅ GA	—	Multi-Region ACID; TrueTime-basiertes Commit-Timestamp
Transaction Savepoints	transaction/	✅ GA	—	createSavepoint/rollbackToSavepoint/releaseSavepoint
OCC Mode	transaction/	✅ GA	—	getEntityVersion/optimisticPut/optimisticErase; per-Entity Versions
Bulk Transaction API	transaction/	✅ GA	—	bulkPutEntities/bulkEraseEntities; kein Per-Row-Overhead
Transaction Explain	transaction/	✅ GA	—	explain(): Locks, Write-Set, MVCC-Version-Chain

Cache-Modul (v1.x)

Feature	Modul	Status	Tests	Beschreibung
AdaptiveQueryCache (L1/L2/L3)	cache/	✅ GA	—	BoundedLRU (L1) + zstd/lz4 (L2) + RocksDB (L3); Adaptive TTL
SemanticCache	cache/	✅ GA	—	SHA-256 Fingerprint + Cosine Similarity; semantische Äquivalenz
GDPR-aware Invalidation	cache/	✅ GA	—	invalidatePII() (Art. 17); UUID-basierte PII-Verknüpfung
Tenant-Isolation + Quotas	cache/	✅ GA	—	per_tenant_max_bytes; cross-tenant Zugriff = nullopt
Circuit Breaker (L3)	cache/	✅ GA	—	CLOSED/OPEN/HALF_OPEN; cb_failure_threshold konfigurierbar
Cache Warmup	cache/	✅ GA	—	warmupFromLog()/exportSnapshot(); Bulk-Vorausladen
Rate Limiting	cache/	✅ GA	—	Token Bucket; max_requests_per_second konfigurierbar

Observability-Modul (v1.x)

Feature	Modul	Status	Tests	Beschreibung
MetricsCollector (Singleton)	observability/	✅ GA	—	Prometheus Text-Format; 40+ Metriken (Query/Cache/Shard/Security/System)
QueryProfiler	observability/	✅ GA	—	Per-Phase Timing (parse/optimize/execute); Index-Usage; Empfehlungen
StorageProfiler	observability/	✅ GA	—	RocksDB-Stats; Write/Read-Amplification; Cache Hit Rates
PerformanceAnalyzer	observability/	✅ GA	—	Automatische Issue-Erkennung; INFO/WARNING/CRITICAL; Empfehlungen
AlertingEngine (Alertmanager)	observability/	✅ GA	—	Rule-based Alerting; Alertmanager Integration; PagerDuty/Slack
Distributed Tracing	observability/	✅ GA	—	OpenTelemetry-kompatibel; Span-Kontext-Propagierung
Kubernetes Health Probes	observability/	✅ GA	—	/ready + /live Endpunkte

Exporters-Modul (v1.x)

Feature	Modul	Status	Tests	Beschreibung
JsonlExporter	exporters/	✅ GA	—	JSONL; Instruction-Tuning: Alpaca/ShareGPT/ChatML/OpenAI
ParquetExporter	exporters/	✅ GA	—	Parquet + konfigurierbare Arrow-Schema; Kompression
ArrowIpcExporter	exporters/	✅ GA	—	Zero-Copy Apache Arrow IPC Pipeline
HuggingFaceExporter	exporters/	✅ GA	—	HF Hub Direct Upload; dataset_info.json + README-Karte
StreamingExporter	exporters/	✅ GA	—	Cursor-basiertes Streaming; Fortschritts-Callback + ETA
IncrementalExporter	exporters/	✅ GA	—	Delta-Export mit Watermark + Checkpoint
AQL Predicate Filter	exporters/	✅ GA	—	AQL-basiertes Filtern bei Export
AES-256-GCM Export Encryption	exporters/	✅ GA	—	Verschlüsselter Export; PolicyEngine-Authorization
PII Detection + Redaction	exporters/	✅ GA	—	Automatische PII-Erkennung + Redaktion vor Export

Implementierungsstatus — Process Module (v1.0)

Serialisierer & Import

Feature	Modul	Status	Tests	Beschreibung
BpmnSerializer::importXml()	process/	✅ GA	BpmnSerializerTest	Zustandsbasierter XML-Tokenizer; 10 MiB-Guard; bpmn:-Namespace-Strip; tolerant
BpmnSerializer::exportXml()	process/	✅ GA	BpmnSerializerTest	Volle BPMN 2.0-Fidelity; OMG-konforme Ausgabe
BpmnSerializer::importFile()	process/	✅ GA	—	Datei-I/O-Wrapper
BpmnSerializer::exportFromJson()	process/	✅ GA	—	Export aus normalisiertem JSON-Graph
EpkSerializer::importText()	process/	✅ GA	EpkSerializerTest	EPK-Text-Notation + JSON-Array-Format
EpkSerializer::exportText()	process/	✅ GA	EpkSerializerTest	Round-trip-fähig
EpkSerializer::exportJson()	process/	✅ GA	EpkSerializerTest	Strukturiertes JSON für LLM-Kontext
VccVpbImporter	process/	✅ GA	VccVpbImporterTest	VCC-VPB YAML-Import

ProcessModelManager

Feature	Modul	Status	Tests	Beschreibung
importBpmn()	process/	✅ GA	ProcessModuleTest	BPMN 2.0 → ProcessModelRecord; RocksDB proc:def:<id>
importEpk()	process/	✅ GA	ProcessModuleTest	EPK → ProcessModelRecord
importVccVpb()	process/	✅ GA	ProcessModuleTest	VCC-VPB → ProcessModelRecord
save()/get()/remove()/list()	process/	✅ GA	ProcessModuleTest	Versioniertes CRUD
deployToEngine()	process/	✅ GA	ProcessModuleTest	Registriert Modell mit ProcessGraphManager
exportBpmn()/exportEpk()	process/	✅ GA	ProcessModuleTest	Rückexport in Originalformat
generateLlmDescriptor()	process/	✅ GA	LlmDescriptorTest	JSON-Schema für LLM-System-Prompts

ProcessLinker

Feature	Modul	Status	Tests	Beschreibung
attachObject()	process/	✅ GA	ProcessModuleTest	8 ProcessLinkType-Werte; node-scoped optional
detachObject()	process/	✅ GA	ProcessModuleTest	Hard-Delete (kein Tombstone)
getAttachments() / getNodeAttachments()	process/	✅ GA	ProcessModuleTest	Mit optionalem Typ-Filter
findInstancesWithObject()	process/	✅ GA	ProcessModuleTest	Sekundärindex proc:obj_idx:
linkProcesses() / getLinks()	process/	✅ GA	ProcessModuleTest	Typisierte Prozess-zu-Prozess-Links
registerRequiredDocument()	process/	✅ GA	ProcessModuleTest	Pflichtdokumentregistrierung pro Modell-Knoten
getMissingDocuments()	process/	✅ GA	ProcessModuleTest	Prüft welche Pflichtdokumente fehlen

ProcessGraphRag

Feature	Modul	Status	Tests	Beschreibung
retrieve()	process/	✅ GA	ProcessGraphRagTest	Vollständiger Graph-RAG-Kontext; DE/EN
retrieveForNode()	process/	✅ GA	ProcessGraphRagTest	Knoten-scoped RAG
summarizeVerwaltungsvorgang()	process/	✅ GA	ProcessGraphRagTest	JSON-Zusammenfassung für UI/API
checkCompliance()	process/	✅ GA	ProcessGraphRagTest	Compliance-Check (Dokumente, SLA, Status)
findSimilarCases()	process/	✅ GA	ProcessGraphRagTest	Cosine/Jaccard-Ähnlichkeit; konfigurierbarer Threshold
buildAdminProcessingPrompt()	process/	✅ GA	ProcessGraphRagTest	Verwaltungsspezifischer LLM-Prompt (DE)

LlmProcessDescriptor

Feature	Modul	Status	Tests	Beschreibung
generateDescriptor()	process/	✅ GA	LlmDescriptorTest	JSON-Schema mit nodes/edges/compliance/sla
generateSystemPrompt()	process/	✅ GA	LlmDescriptorTest	System-Prompt DE/EN
buildConformancePrompt()	process/	✅ GA	LlmDescriptorTest	Conformance-Checking-Prompt

---

Version History:

• 1.4.0-alpha (2026-01-06): 25 neue Alpha-Features (LLM, Performance, HA, Monitoring, Protocol)
• 1.3.0 (2025-12-30): Umfassendes Update mit allen Features bis Dez 2025

Implementierungsstatus — Training, CDC & Updates (2026-04-12)

Training-Modul (v1.5)

Feature	Header	Status	Beschreibung
TrainingPipeline::run()	training/training_pipeline.h	✅ GA	6-stufige End-to-End LoRA-Pipeline aus ThemisDB-Collections
IncrementalLoRATrainer::train()	training/incremental_lora_trainer.h	✅ GA	INITIAL/RESUME/CONTINUE-Modes, INT8/INT4 Quantisierung
IncrementalLoRATrainer::resumeFromCheckpoint()	training/incremental_lora_trainer.h	✅ GA	SHA-256-verifiziertes Laden, Auto-Rollback
AutoLabeler::label()	training/auto_labeler.h	✅ GA	LLM-Annotation + Keyword-Extraktion, Konfidenz-Filter
KnowledgeGraphEnricher	training/knowledge_graph_enricher.h	✅ GA	KG-Kontext je Sample aus ThemisDB-Graph
LoraDataSelection	training/lora_data_selection.h	✅ GA	Aktives Lernen / Uncertainty Sampling
LoRACheckpointManager::save()/resume()	training/lora_checkpoint_manager.h	✅ GA	Atomares Speichern, SHA-256, Auto-Rollback, JSON-Manifest
ProvenanceTracker::record()	training/provenance_tracker.h	✅ GA	Lineage-Graph für Gewichts-Änderungen
ModalityParser	training/modality_parser.h	✅ GA	Multimodale Input-Normalisierung (Text/Bild/Audio)
LoRAAdapterMerger	training/lora_adapter_merger.h	✅ GA	Mergen mehrerer LoRA-Adapter (gewichtete Linearkombination)

CDC-Modul (v1.x)

Feature	Header	Status	Beschreibung
Changefeed::listEvents()	cdc/changefeed.h	✅ GA	Sequenz-basiert, Key-Präfix-Filter, Long-Polling
Changefeed::subscribe()	cdc/changefeed.h	✅ GA	Push-Subscription mit SubscriptionFilter + cancel()
Changefeed::compactByKey()	cdc/changefeed.h	✅ GA	Manuelle Kompaktierung, removed_count + bytes_freed
Changefeed::redactByKeyPrefix()	cdc/changefeed.h	✅ GA	GDPR-Redaktion im Event-Log
Changefeed::watermarks()	cdc/changefeed.h	✅ GA	min_seq, max_seq, consumer_lag
CdcMaterializedView	cdc/cdc_materialized_view.h	✅ GA	Live-Sync MV via Changefeed
CdcAdmin	cdc/cdc_admin.h	✅ GA	Feed-Verwaltung, Reset, Export
ConsumerGroup	cdc/consumer_group.h	✅ GA	Multi-Consumer-Offset-Management
CrossCollectionStream	cdc/cross_collection_stream.h	✅ GA	Joins über mehrere Changefeeds

Updates-Modul (v1.x)

Feature	Header	Status	Beschreibung
ClusterUpdateManager::update()	updates/cluster_update_manager.h	✅ GA	Rolling-Update mit DRAIN/APPLY/HEALTHCHECK/ROLLBACK
CanaryRollout::deploy()/promote()/rollback()	updates/canary_rollout.h	✅ GA	Prozentbasierter Canary, automatischer Rollback bei Error-Rate
HotReloadEngine::reloadConfig()/reloadPlugin()	updates/hot_reload_engine.h	✅ GA	Zero-Downtime Konfigurations- und Plugin-Reload
InPlaceSchemaMigrator::execute()	updates/in_place_schema_migrator.h	✅ GA	Online-Schema-Migration ohne Downtime
BlueGreenDeployment::cutover()/rollback()	updates/blue_green_deployment.h	✅ GA	Instant Traffic-Switch Blue↔Green
DeltaUpdateEngine	updates/delta_update_engine.h	✅ GA	Inkrementelle Patch-Anwendung auf Datenbankebene
DependencyResolver	updates/dependency_resolver.h	✅ GA	Update-Reihenfolge basierend auf Abhängigkeitsgraph
ManifestDatabase	updates/manifest_database.h	✅ GA	Versionierter Software-Manifest-Store
CoordinatedUpdateManager	updates/coordinated_update_manager.h	✅ GA	Multi-Cluster koordinierte Updates mit Quorum-Fencing

Implementierungsstatus — Auth, BiTemporal, Storage, LLM-Infrastruktur (2026-04-13)

Auth-Modul C++ API (v1.x)

Feature	Header	Status	Beschreibung
JWTValidator::validate()	auth/jwt_validator.h	✅ GA	RS256/ES256/EdDSA, Blacklist, KID-Revokation
OAuthPKCEFlow::exchangeCode()	auth/oauth_pkce_flow.h	✅ GA	RFC 7636 PKCE, Authorization-Code-Flow
SAMLAuthenticator::validateResponse()	auth/saml_authenticator.h	✅ GA	SAML 2.0 SP, Attribut-Mapping, IdP-Zertifikat
WebAuthnAuthenticator::finishRegistration/Authentication()	auth/webauthn_authenticator.h	✅ GA	FIDO2/WebAuthn, PassKey-Support
LDAPAuthenticator::authenticate()	auth/ldap_authenticator.h	✅ GA	LDAPS, Connection-Pool, Group-Lookup
MFAAuthenticator::validateTOTP()	auth/mfa_authenticator.h	✅ GA	TOTP RFC 6238, Recovery Codes, Enrollment
SessionManager::validateAndRefresh()	auth/session_manager.h	✅ GA	Idle/Absolute-Timeout, Max-Sessions-Per-User
PasswordPolicy::validate()	auth/password_policy.h	✅ GA	Entropie-Scoring, HaveIBeenPwned-Check
FederatedIdentityManager::validateToken()	auth/federated_identity_manager.h	✅ GA	OIDC Multi-Realm, Token-Exchange
JwksValidator	auth/jwks_validator.h	✅ GA	JWKS-Endpoint, Auto-Refresh
JwtKeyRotationManager	auth/jwt_key_rotation_manager.h	✅ GA	Key-Rotation ohne Downtime
OidcProvider	auth/oidc_provider.h	✅ GA	OIDC Discovery, UserInfo-Endpoint
ApiKeyAuthenticator	auth/api_key_authenticator.h	✅ GA	Scoped API-Keys, Constant-Time-Vergleich
AuthRateLimiter	auth/auth_rate_limiter.h	✅ GA	Sliding-Window, Per-User/Per-IP
MtlsAuthenticator	auth/mtls_authenticator.h	✅ GA	Client-Zertifikat-Authentifizierung

Bi-Temporal-Modul (v1.x)

Feature	Header	Status	Beschreibung
BiTemporalTable::insertWithValidTime()	temporal/bi_temporal.h	✅ GA	SQL:2011 Bi-Temporale Einfügung
BiTemporalTable::queryAsOf()	temporal/bi_temporal.h	✅ GA	AS OF Abfrage (System-Zeit + Gültigkeits-Zeit)
BiTemporalTable::detectGaps()	temporal/bi_temporal.h	✅ GA	Lücken-Erkennung im Gültigkeits-Zeitraum
TemporalForeignKey::validate()	temporal/bi_temporal.h	✅ GA	Referentielle Integrität über Zeiträume
BiTemporalJoin	temporal/bitemporal_join.h	✅ GA	SEQUENCED/NON_SEQUENCED/CURRENT Join
TemporalQueryEngine::execute()	temporal/temporal_query_engine.h	✅ GA	AS OF / FROM-TO / CONTAINED IN Abfragen
TemporalSnapshotManager	temporal/snapshot_manager.h	✅ GA	Snapshot-Handle für PITR
SystemVersionedTable	temporal/system_versioned_table.h	✅ GA	Automatische System-Zeit-Verwaltung
IntervalTreeIndex	temporal/interval_tree_index.h	✅ GA	Effiziente Overlap-Suche
TemporalAggregator	temporal/temporal_aggregator.h	✅ GA	Zeitraum-aggregierte Aggregate (SUM/AVG/COUNT)

Storage/Backup-Modul C++ API (v1.x)

Feature	Header	Status	Beschreibung
BackupManager::createFullBackup()	storage/backup_manager.h	✅ GA	Full Backup, AES-256-GCM, RAID-aware
BackupManager::createIncrementalBackup()	storage/backup_manager.h	✅ GA	Inkrementell seit letztem Backup
BackupManager::createDifferentialBackup()	storage/backup_manager.h	✅ GA	Differenziell seit letztem Full-Backup
BackupManager::restoreFromBackup()	storage/backup_manager.h	✅ GA	RAID-Shard-Rekonstruktion
BackupManager::archiveWAL()	storage/backup_manager.h	✅ GA	WAL-Archivierung für PITR
PITRManager::restore()	storage/pitr_manager.h	✅ GA	Point-in-Time Recovery, Fortschritts-Callback
PITRManager::previewRestore()	storage/pitr_manager.h	✅ GA	Dry-Run ohne Änderungen
TieredStorageManager::put()/get()	storage/tiered_storage.h	✅ GA	HOT/WARM/COLD transparent, Auto-Promotion
AdaptiveCompactionManager	storage/adaptive_compaction.h	✅ GA	Write/Read-Amplification-basierte Strategie
BlobStorageManager	storage/blob_storage_manager.h	✅ GA	Multi-Backend (FS/GCS/S3), Zero-Copy
ErasureCodingBackend	storage/erasure_coding_backend.h	✅ GA	Reed-Solomon, konfigurierbare Parität

LLM-Infrastruktur (v1.x)

Feature	Header	Status	Beschreibung
PagedKVCache::store()/sharePrefix()	llm/paged_kv_cache.h	✅ GA	vLLM Paged Attention, Prefix-Sharing
ContinuousBatchScheduler::step()	llm/continuous_batch_scheduler.h	✅ GA	Continuous Batching, Preemption, Priority
SpeculativeDecoder::decode()	llm/speculative_decoder.h	✅ GA	Draft-Model, 5-Token-Lookahead, Acceptance-Threshold
OpenAICompatAdapter	llm/openai_compat_adapter.h	✅ GA	/v1/chat/completions + Streaming
LoRARouter::route()	llm/lora_router.h	✅ GA	A/B-Testing, Canary-Rollout, Fallback
AdapterRegistry::registerAdapter()	llm/adapter_registry.h	✅ GA	Versionierter Adapter-Store, Provenance
ModelRouter::route()	llm/model_router.h	✅ GA	Rule-based Routing (Kontext/Tenant/Custom)
AdapterLoadBalancer	llm/adapter_load_balancer.h	✅ GA	GPU-Placement, JIT-Eviction, Migration
AdapterDeploymentManager	llm/adapter_deployment_manager.h	✅ GA	Blue/Green, Canary für Adapter
LlamaResourceManager	llm/llama_resource_manager.h	✅ GA	llama.cpp Ressourcen-Lifecycle
GpuMemoryManager	llm/gpu_memory_manager.h	✅ GA	VRAM-Budget, OOM-Safe-Fail
EmbeddedLLM	llm/embedded_llm.h	✅ GA	In-Process LLM ohne Server

Implementierungsstatus — RAG Advanced, Sharding C++ API, Query/Index Engine (2026-04-13)

RAG-Modul v2 — Erweiterte Pipeline-Komponenten

Feature	Header	Status	Beschreibung
AgenticRAG::run()	rag/agentic_rag.h	✅ GA	Iterativer Retrieve-then-Reason Agent (max_iterations, quality_threshold, cancel)
AgenticRAGFactory	rag/agentic_rag.h	✅ GA	Factory für Standard- und Custom-Konfigurationen
MultiStepRAGOrchestrator::run()	rag/multi_step_rag.h	✅ GA	Decompose-then-Retrieve (max_steps, LLM-Decomposition, MergeStrategy)
MultiModalRAG::query()	rag/multimodal_rag.h	✅ GA	TEXT/IMAGE/TABLE/CODE/AUDIO; OCR-Support
RAGContextAssembler::assemble()	rag/rag_context_assembler.h	✅ GA	Token-Budget-Management, Dedup, Ordering
DistributedRAGEvaluator::evaluate()	rag/distributed_rag_evaluator.h	✅ GA	Parallele Multi-Judge-Evaluierung, Quorum, AggregationStrategy
DistributedEvaluatorFactory	rag/distributed_rag_evaluator.h	✅ GA	Builder für Distributed-Judge-Cluster

Sharding-Modul C++ API (v1.x)

Feature	Header	Status	Beschreibung
AdaptiveShardRouter::route()	sharding/adaptive_shard_router.h	✅ GA	Lernender Router mit Exploration/Exploitation, Feedback-Loop
ConsistentHashRing::getShardForKey()	sharding/consistent_hash.h	✅ GA	Virtueller Knoten-Ring, 150 vNodes/Shard, getNShardsForKey
CrossShardTransaction::commit()	sharding/cross_shard_transaction.h	✅ GA	2PC + WAL, IsolationLevel SERIALIZABLE/SNAPSHOT/READ_COMMITTED
EpochFencingManager::checkFence()	sharding/epoch_fencing.h	✅ GA	Stale-Leader-Schutz, Lease-Management, advanceEpoch
HotShardSplitPolicy	sharding/auto_rebalancer.h	✅ GA	ML-prädiktive Shard-Aufteilung, QPS-Threshold
DataMigrator::migrate()	sharding/data_migrator.h	✅ GA	Live-Migration mit Dual-Write, Verifikation, Rate-Limiting
AutoRecoveryManager	sharding/auto_recovery_manager.h	✅ GA	Automatische Shard-Recovery nach Ausfall
GossipConfigManager	sharding/gossip_config_manager.h	✅ GA	Gossip-basierte Konfigurationsverteilung

Query Engine C++ API (v1.x)

Feature	Header	Status	Beschreibung
AdaptiveQueryStats::recordExecution()	query/adaptive_optimizer.h	✅ GA	Query-Statistiken akkumulieren, Kardinalitätsfehler erkennen
AdaptivePlanSelector::selectPlan()	query/adaptive_optimizer.h	✅ GA	KEEP/SWITCH/PARALLEL_TEST basierend auf Feedback
DistributedQueryCostModel	query/adaptive_optimizer.h	✅ GA	Cross-Shard-Join-Kosten, Partition Pruning
MultiIndexOptimizer	query/adaptive_optimizer.h	✅ GA	Optimale Index-Kombination für komplexe Prädikate
AdaptiveJoin::executeJoin()	query/adaptive_join.h	✅ GA	HASH/SORT_MERGE/INDEX/NESTED_LOOP/GRACE_HASH, Runtime-Wahl
ApproximateCountDistinct (HyperLogLog)	query/approximate_aggregator.h	✅ GA	±2 % Fehler, merge für verteilte Aggregation
ApproximatePercentile (T-Digest)	query/approximate_aggregator.h	✅ GA	p50/p95/p99, merge-fähig

Index-Modul C++ API (v1.x)

Feature	Header	Status	Beschreibung
IndexManager	index/index_manager.h	✅ GA	Koordiniert Vector/Secondary/Graph-IndexManager
InvertedIndex::search()	index/inverted_index.h	✅ GA	BM25, Stemming, Stopwords, RocksDB-backed
AdvancedVectorIndex	index/advanced_vector_index.h	✅ GA	AUTO-Type (HNSW/IVF/SQ/PQ), WorkloadType-aware
DistributedVectorIndex	index/distributed_vector_index.h	✅ GA	Scatter-Gather K-NN, ConsistentHash, Replikation
BinaryQuantizer	index/binary_quantizer.h	✅ GA	1-bit Quantisierung, FAISS-Integration
IndexSuggestionEngine::recommend()	index/adaptive_index.h	✅ GA	Auto-Empfehlung basierend auf Query-Patterns + Selektivität
QueryPatternTracker	index/adaptive_index.h	✅ GA	Scan-Latenz, Rows-scanned, Cache-Miss aufzeichnen
SelectivityAnalyzer	index/adaptive_index.h	✅ GA	Distinct-Values, Histogramm, Selektivitäts-Score

Implementierungsstatus — Server/API, Metadata, Performance, Plugins, UserStorage (2026-04-13 session 4)

Server/API-Modul (v1.x)

Feature	Header	Status	Beschreibung
APIGateway::registerHandler()	server/api_gateway.h	✅ GA	Einheitlicher HTTP-Einstiegspunkt; Route-Registrierung, Load-Shedding, Deprecation-Header
AdaptiveRateLimiter::allowRequest()	server/adaptive_rate_limiter.h	✅ GA	Feedback-basiertes Rate-Limiting, Burst-Toleranz, Overload-Threshold
AuthMiddleware::authenticate()	server/auth_middleware.h	✅ GA	JWT/GSSAPI/API-Key/USB-Token; Scope-Prüfung
DistributedGateway / ConsistentHashRing	server/distributed_gateway.h	✅ GA	Cluster-Routing, Circuit-Breaker, Failover, 200 vNodes
AsyncJobApiHandler / AsyncJobRegistry	server/async_job_api_handler.h	✅ GA	Async-Job-Queue (PENDING/RUNNING/COMPLETED/FAILED), prune()
GrpcApiServer::start()	api/grpc_server.h	✅ GA	gRPC Port 8771, TLS, max-message-size, registerService
GraphQLSchemaBuilder::build()	api/graphql_schema_builder.h	✅ GA	Type/Query/Mutation-Registrierung, Schema-Validierung
IWebSocketFrameCallback	api/websocket_handler.h	✅ GA	WebSocket-Frame-Callbacks (Text/Binary/Ping/Close), CloseCode

Metadata/Schema-Modul (v1.x)

Feature	Header	Status	Beschreibung
SchemaManager::createTable()	metadata/schema_manager.h	✅ GA	Tabellen/Relationships/Indexes; AdaptiveTTL-Konfiguration
InformationSchema::getTables()	metadata/information_schema.h	✅ GA	SQL-kompatible IS: tables/columns/statistics/referential_constraints
SchemaVersionManager::persistChange()	metadata/schema_version_manager.h	✅ GA	WAL-backed Migrations; rollback; AuditLog-Integration
DistributedMetadataCatalog::publishSchema()	metadata/distributed_catalog.h	✅ GA	Cluster-weite Schema-Verteilung via Consensus
SchemaConsistencyChecker::check()	metadata/schema_consistency_checker.h	✅ GA	Schema↔Daten-Konsistenz; Auto-Repair; dry_run
IndexRecommender	metadata/index_recommender.h	✅ GA	Empfiehlt Indexes basierend auf Query-Statistiken
SchemaAuditLog	metadata/schema_audit_log.h	✅ GA	Append-only Schema-Änderungsprotokoll
ERDiagramExporter	metadata/er_diagram_exporter.h	✅ GA	ER-Diagramm-Export (Mermaid/PlantUML)

Performance-Internals (v1.x)

Feature	Header	Status	Beschreibung
AdaptiveQueryCompiler::compile()	performance/adaptive_query_compiler.h	✅ GA	JIT-ähnliche Query-Kompilierung; Hot-Query-Cache
IntelligentPrefetcher::predict_next()	performance/intelligent_prefetcher.h	✅ GA	ML-basiertes Prefetching (SEQUENTIAL/STRIDED/RANDOM/POINTER_CHASE); Hardware-Prefetch
WorkloadPredictor::forecast()	performance/workload_predictor.h	✅ GA	LSTM/EMA Lastvorhersage; ScaleDirection (UP/DOWN/STABLE); Replikationsempfehlung
HardwareAccelerator::execute()	performance/hardware_accelerator.h	✅ GA	CPU/CUDA/OpenCL/NPU/FPGA Dispatch; OperatorType-Auswahl
LockFreeRingBuffer<T>	performance/lockfree_metrics_buffer.h	✅ GA	SPSC Lock-Free FIFO; tryPush/tryPop; dropped_count
WorkloadPredictor::getScaleRecommendation()	performance/workload_predictor.h	✅ GA	Konkrete Replikations-Empfehlung mit Konfidenz
WorkloadAdaptiveOptimizer::classify_workload()	performance/workload_adaptive_optimizer.h	✅ GA	Heuristische Klassifikation (OLTP/OLAP/MIXED/GRAPH/VECTOR/TIMESERIES); 512-Query Rolling Window
WorkloadAdaptiveOptimizer::get_strategy()	performance/workload_adaptive_optimizer.h	✅ GA	Per-Workload-Strategie: JIT, Thread-Pool, Cache-MB, Join-Algorithmus, Index-Typ; Predictive Scaling
WorkloadAdaptiveOptimizer::enable_auto_adapt()	performance/workload_adaptive_optimizer.h	✅ GA	Hintergrund-Adaptionsloop (konfigurierbares Intervall); AdaptationCallback

Plugin-System (v1.x)

Feature	Header	Status	Beschreibung
PluginManager::loadPlugin()	plugins/plugin_manager.h	✅ GA	Signatur-Verifikation, Hot-Reload, PluginReloadPhase-Callbacks
WasmHostAPI	plugins/wasm_host_api.h	✅ GA	WASM-Plugins (Wasmtime/Wasmer/Wasm3); Host-Function-ABI
PluginHealthMonitor::start()	plugins/plugin_health_monitor.h	✅ GA	Auto-Disable + Recovery; MonitoringEvent-Callbacks; Health-Score
SignedPluginRepository	plugins/signed_plugin_repository.h	✅ GA	OCI-Registry-Integration für signierte Plugin-Pakete
SelfHealingPlugin	plugins/self_healing_plugin.h	✅ GA	Automatische Neustart-Logik bei Crash
PluginDependencyResolver	plugins/plugin_dependency_resolver.h	✅ GA	Topologische Sortierung + Zyklenerkennung für Plugin-Deps

User Storage Encrypted (v1.x)

Feature	Header	Status	Beschreibung
MultiLevelEncryptedStorage::mount()	user_storage_encrypted/multi_level_storage.hpp	✅ GA	FUSE-basierter Multi-Level-Verschlüsselungsspeicher; SecurityLevel LOW/MEDIUM/HIGH
KeyRotationScheduler::scheduleRotation()	user_storage_encrypted/key_rotation_scheduler.hpp	✅ GA	Geplante + manuelle Schlüsselrotation; IRotationStore-Persistenz
GocryptfsBackend::mount()	user_storage_encrypted/gocryptfs_backend.hpp	✅ GA	Gocryptfs FUSE-Backend; mount/unmount/isMounted
Argon2idKeyDerivationService::deriveKey()	user_storage_encrypted/key_derivation_service.hpp	✅ GA	Argon2id KDF (OWASP-Empfehlung); memory/iterations/parallelism
reconcileStaleMounts()	user_storage_encrypted/multi_level_storage.hpp	✅ GA	Aufräumen verwaister FUSE-Mounts beim Systemstart

Implementierungsstatus — Geo Extended, Timeseries Internals, AQL Advanced, Scheduler (2026-04-13 session 5)

Geo-Modul Erweiterungen (v2.x)

Feature	Header	Status	Beschreibung
GeoRTree::bulkLoad() / queryBBox() / queryKNN()	geo/geo_rtree.h	✅ GA	R*-Tree Spatial Index; Bulk-Load, BBox-Query, K-NN
SpatialJoinIterator::advance()	geo/spatial_join.h	✅ GA	Lazy Iterator; INTERSECTS/WITHIN_DISTANCE/CONTAINS/OVERLAPS
TileServer::getTile()	geo/tile_server.h	✅ GA	Mapbox Vector Tiles (MVT/Protobuf); Layer-Config; Geometrie-Vereinfachung
geoDbscan() / geoKMeans()	geo/geo_clustering.h	✅ GA	DBSCAN (eps_m, min_points) + K-Means (k, max_iter); Noise-Points
RasterGrid::sample() / generateHeatmap()	geo/raster.h	✅ GA	Bilinear-Sampling, Gaussian-Heatmap, normalize

Timeseries-Internals (v1.9.x)

Feature	Header	Status	Beschreibung
TimeSeriesStore::put() / query() / aggregate()	timeseries/timeseries.h	✅ GA	Core-KV-Store; Range-Query + Window-Aggregation
Hypertable::insertBatch() / compressChunk()	timeseries/hypertable.h	✅ GA	Auto-Partitionierung nach Zeit; Gorilla+Zstd Chunk-Kompression
ContinuousAggWatermarkStore / DistributedAggregateCoordinator	timeseries/continuous_agg.h	✅ GA	Inkrementelle Materialisierung; Rollup-Hierarchien; Watermark
BitWriter / BitReader (Gorilla)	timeseries/gorilla.h	✅ GA	Delta-of-Delta Timestamp + XOR Value Encoding
DownsamplingPipeline::addPolicy()	timeseries/downsampling.h	✅ GA	Mehrstufiges Downsampling; AggFunc-Auswahl; Retention
RetentionManager::startAsync()	timeseries/retention.h	✅ GA	Soft+Hard Delete; StagedDeletion; Audit-Trail
TsStreamCursor::nextBatch()	timeseries/ts_stream_cursor.h	✅ GA	Batch-Streaming-Iterator über Zeitreihen
PrometheusRemoteWriteHandler::handle()	timeseries/prometheus_remote_write.h	✅ GA	Prometheus Remote-Write-Endpoint; Grafana-kompatibles Remote-Read

AQL Advanced API (v1.x)

Feature	Header	Status	Beschreibung
AQLQueryBuilder::from() / build()	aql/aql_query_builder.h	✅ GA	Fluent Builder; FROM/FILTER/SORT/LIMIT/JOIN/SELECT
AQLQueryValidator::validate()	aql/aql_query_validator.h	✅ GA	Syntaktisch + semantisch; ERROR/WARNING/INFO; Line/Column
AQLOptimizerAdvisor::analyze()	aql/aql_optimizer_advisor.h	✅ GA	Index-Empfehlungen; Rewrite-Vorschläge; estimated_speedup
AQLConversationContext::ask()	aql/aql_conversation_context.h	✅ GA	Multi-Turn NL→AQL; Schema-Kontext; Auto-Correct
AQLMigrationAssistant::migrate()	aql/aql_migration_assistant.h	✅ GA	SQL→AQL Migrationshilfe; BLOCKING/WARNING Issues
AQLAgent::run()	aql/aql_agent.h	✅ GA	LLM-gesteuerter DB-Agent; Tool-Registrierung; Reasoning-Steps

Scheduler-Internals (v1.x)

Feature	Header	Status	Beschreibung
TaskScheduler::addTask()	scheduler/task_scheduler.h	✅ GA	CRON/INTERVAL/CDC/MANUAL/WEBHOOK; 5/6-Feld, Timezone, CDC-Trigger
DistributedTaskCoordinator::start()	scheduler/distributed_task_coordinator.h	✅ GA	Leader-Election; One-Runner-Per-Cluster; LeaderChange-Callback
TaskAnomalyDetector::analyzeTask()	scheduler/task_anomaly_detector.h	✅ GA	Frequenz/Pattern/Resource/Failure-Rate-Erkennung; Callback
TaskAuditManager::query() / exportTo()	scheduler/task_audit_manager.h	✅ GA	Vollständiger Audit-Trail; JSON/CSV/Parquet-Export; 90-Tage-Retention
HybridRetentionManager::startAsync()	scheduler/hybrid_retention_manager.h	✅ GA	3-Stufen: Gorilla (0–7d) + Downsampling (7–365d) + Tages-Agg (>365d)

Implementierungsstatus — LLM Advanced AI, Replication, Content Module, Acceleration (2026-04-13 session 6)

LLM Advanced AI (v2.x)

Feature	Header	Status	Beschreibung
AiOrchestrator::run()	llm/ai_orchestrator.h	✅ GA	Multi-Mode Pipeline: RAG/AGENTIC/CRITIQUE; ModeSpec, BudgetSpec, SafetySpec, JudgeSpec
AsyncInferenceEngine::submitAsync()	llm/async_inference_engine.h	✅ GA	Streaming Token-Callback; DROP_OLDEST/BLOCK Backpressure; Dedup-Cache
InferenceEngineEnhanced	llm/inference_engine_enhanced.h	✅ GA	Multi-Model LB (ROUND_ROBIN/LEAST_LOADED/FASTEST); Circuit Breaker; ModelResourceQuota
InlineTrainingEngine::stepAsync()	llm/inline_training_engine.h	✅ GA	On-the-Fly LoRA Fine-Tuning; AdamW/LION Optimizer; Cosine LR Scheduler
ConstitutionalReasoningEngine::reason()	llm/constitutional_reasoning_engine.h	✅ GA	Prinzipienbasiert; max_revision_rounds; hard_block; HHH-Prinzipien
EthicsAwareConfidenceDetector::evaluate()	llm/ethics_aware_confidence_detector.h	✅ GA	Bias-Kategorien; Token-Konfidenz; Ethics + Uncertainty Combined Score

Replication Multi-Master & CRDT (v1.x)

Feature	Header	Status	Beschreibung
ReplicationManager::addReplica()	replication/replication_manager.h	✅ GA	PRIMARY/SECONDARY/ARBITER; SYNC/ASYNC/SEMI_SYNC; Lag-Monitoring
VectorClock::happensBefore() / isConcurrent()	replication/multi_master_replication.h	✅ GA	Kausalordnung; Nebenläufigkeitserkennung; merge()
HybridLogicalClock::now()	replication/multi_master_replication.h	✅ GA	Wall + Logical Timestamp; TrueTime-ähnlich
ConflictResolver::resolve()	replication/multi_master_replication.h	✅ GA	LWW/FWW/MERGE/CUSTOM; CONCURRENT_WRITE/DELETE_UPDATE/SCHEMA_CONFLICT
GrowOnlyCounter / PNCounter / MVRegister (CRDT)	replication/crdt_types.h	✅ GA	Kommutative CRDT-Merge; G-Counter, PN-Counter, Multi-Value Register
ReplicationSlot::advance() / drop()	replication/replication_slot.h	✅ GA	Persistent WAL-Slots; LSN-Fortschritt; PAUSE/RESUME

Content-Modul (v1.x)

Feature	Header	Status	Beschreibung
ContentManager::ingestRawBlob() Pipeline 10 Stages	content/content_manager.h	✅ GA	MIME-Auto, Malware-Scan, pHash-Dedup, OCR, Chunking, Embed, LLM-Tagging
EmbeddingPipeline::embed()	content/embedding_pipeline.h	✅ GA	Batch-Embedding; normalize; Failure-Callback; Device-Auswahl
AsyncIngestionWorker::enqueue()	content/async_ingestion_worker.h	✅ GA	4 Worker-Threads; QUEUED/RUNNING/FAILED/RETRYING; Priority-Queue
ContentSecurity::scan()	content/content_security.h	✅ GA	Malware/ZIP-Bomb/Abuse; sanitize_error_messages; block_on_malware
IContentProcessor Plugin-Interface	content/content_processor.h	✅ GA	Eigene MIME-Typen; ExtractionResult (text/language/geo_data/media_data)
ContentPolicy MIME-Whitelist/Blacklist	content/content_policy.h	✅ GA	allowed_mime_types; enable_deduplication; ocr_enabled; isAllowed()

Acceleration-Internals (v1.x)

Feature	Header	Status	Beschreibung
ComputeBackend::executeSimilarityKernel()	acceleration/compute_backend.h	✅ GA	FP32/FP16/BF16/INT8/INT4; CUDA/HIP/Vulkan/OpenCL/DirectX; Health-Status
ANNKernelFallbackDispatcher::search()	acceleration/kernel_fallback_dispatcher.h	✅ GA	Fallback-Kette CUDA→HIP→Vulkan→CPU; RetryPolicy; backend_used
GeoKernelFallbackDispatcher::computeDistancesBatch()	acceleration/kernel_fallback_dispatcher.h	✅ GA	Haversine/Vincenty Batch; GPU-Fallback auf CPU
VecKnnPipeline::search() / searchBatch()	acceleration/vec_knn.h	✅ GA	IVF nprobe + HNSW ef_search; Distance-Cache; AQL-Filter
DeviceManager::getDevices()	acceleration/device_manager.h	✅ GA	Multi-GPU-Inventar; VRAM/Compute-Units; hasGPU(); logDeviceInfo()
CudaBuffer<T> / VulkanBuffer<T> RAII	acceleration/raii/cuda_raii.h	✅ GA	Automatische GPU-Speicherfreigabe; upload/download

---

ThemisDB 1.9.0-beta · Home · Module-Index · GitHub · Issues