Systems

The MinIO server is a single binary that contains many subsystems. This section covers the major internal building blocks. They are organised here as "systems" because most of them live inside cmd/ and don't map cleanly to either an external package boundary or a deployable unit.

The runtime is the only deployable: there is one minio binary that can run as a single-node server or as one node in a distributed cluster. The "applications" lens is therefore not used; instead, the systems below describe the moving parts inside that binary.

At a glance

graph TD
    HTTP[HTTP / API surface] --> OBJ[Object layer]
    OBJ --> EC[Erasure coding]
    OBJ --> META[Metacache & listing]
    OBJ --> LCK[Locking & DSync]
    OBJ --> SCAN[Data scanner]
    OBJ --> HEAL[Healing & MRF]
    OBJ --> REPL[Replication]
    OBJ --> ILM[Lifecycle / Tiering]
    HTTP --> IAM[IAM / STS]
    HTTP --> KMS[KMS / encryption]
    HTTP --> EVT[Events / Notifications]
    HTTP --> METRICS[Metrics & audit]
    OBJ --> NET[Peer / Storage RPC]
    NET --> GRID[Grid RPC]

Subsystems

Subsystem	Brief	Dedicated page
Object API & HTTP	The S3 routing surface, request middleware, and response helpers.	object-api-and-http
Erasure coding	Server pools, sets, encode/decode, xl.meta, on-disk format.	erasure
Distributed locking	Namespace lock, DSync, lock REST.	distributed-locking
Healing	Per-object heal, background heal, MRF, new-disks heal.	healing
Data scanner	Background sweep that drives quotas, ILM, and heal hints.	data-scanner
IAM and STS	Users, groups, policies, identity providers, temporary credentials.	iam
KMS and encryption	SSE-S3, SSE-KMS, SSE-C wiring; KES integration.	kms
Bucket replication	Async per-bucket replication to S3-compatible targets.	replication
Site replication	Cluster-to-cluster mirror of buckets, IAM, lifecycle, encryption.	site-replication
Lifecycle & tiering	Bucket ILM, transitions to warm backends.	lifecycle-and-tiering
Metacache & listing	Streaming list-objects cache.	metacache
Notifications	Bucket event delivery to Kafka, NATS, MQTT, Webhook, ES, etc.	notifications
Metrics, audit, and tracing	Prometheus v2/v3, audit log, request tracing.	metrics-and-audit
Batch jobs	Bulk replicate / expire / key-rotate jobs run by the cluster.	batch-jobs
FTP and SFTP	Embedded FTP/FTPS and SFTP servers.	ftp-and-sftp
Object Lambda	Request-time transforms via external functions.	object-lambda
Admin API	Cluster admin REST surface used by mc admin.	admin-api
Peer & Storage RPC	Inter-node RPC layers (storage REST, peer REST, grid).	peer-and-storage-rpc