Glossary
The acronyms and the recurring technical terms used across these docs, expanded once here.
| Term | Expansion | In short |
|---|---|---|
| AAI | Authentication and Authorization Infrastructure | Swiss higher-education identity federation, operated by SWITCH |
| ACL | Access Control List | Rules defining who may access what |
| BMC / IPMI | Baseboard Management Controller / Intelligent Platform Management Interface | Out-of-band hardware management (console, power, virtual media), independent of the OS. Branded iDRAC by Dell, AST2500 on Gigabyte |
| CE | Community Edition | The free, open-source edition of GitLab |
| CI/CD | Continuous Integration / Continuous Delivery | Automatic test and deployment runs on every code change |
| cluster | — | Several Proxmox nodes managed as one, sharing configuration and able to migrate guests between themselves |
| corosync | — | The communication layer between Proxmox nodes: detects failures and maintains quorum |
| CUDA | Compute Unified Device Architecture | NVIDIA's GPU computing platform |
| DAC | Direct Attach Copper | Copper cable with integrated transceivers, for short server-to-switch links |
| DNS | Domain Name System | Translation of domain names into IP addresses |
| DWPD | Drive Writes Per Day | SSD endurance: how many times per day its full capacity can be rewritten over 5 years |
| FQDN | Fully Qualified Domain Name | Complete domain name (e.g. git.isc.hevs.ch) |
| HA | High Availability | Automatic restart of critical VMs on another node after a failure |
| HBA | Host Bus Adapter | Disk controller with no RAID layer. In IT mode it presents raw disks to the OS — essential for ZFS |
| hot spare | — | A standby disk installed in the machine that ZFS enrols automatically on failure |
| IdP | Identity Provider | Here, SWITCH edu-ID |
| IDS | Intrusion Detection System | Intrusion detection (Suricata on the network, Wazuh on hosts) |
| K8s | Kubernetes | Container orchestrator |
| LXC | Linux Containers | Lightweight system containers (sharing the host kernel), an alternative to VMs in Proxmox |
| NOC | Network Operations Center | Monitoring infrastructure health (metrics, availability, alerts) |
| NVMe | Non-Volatile Memory Express | SSD access protocol over PCIe, far faster than SATA/SAS |
| OCP | Open Compute Project | Network card format (OCP 3.0), mounted on a mezzanine without consuming a PCIe slot |
| OIDC | OpenID Connect | Modern authentication protocol built on OAuth 2.0 — chosen for edu-ID integration |
| PBS | Proxmox Backup Server | Dedicated backup server for Proxmox VMs/containers (deduplicated, incremental, verified) |
| PDM | Proxmox Datacenter Manager | Central console for driving multiple Proxmox hosts/clusters. A separate product from the "Datacenter" node in the PVE interface |
| PDU | Power Distribution Unit | Rack power strip, often remotely switchable and meterable |
| PLP | Power Loss Protection | Capacitors in an enterprise SSD guaranteeing cache flush on power loss. Absent from consumer SSDs — a decisive criterion for ZFS |
| PVE | Proxmox Virtual Environment | The hypervisor: the virtualisation layer installed on each machine |
| QDevice | — | An external tie-breaker vote for a Proxmox cluster, so an even number of nodes (two, typically) can still reach quorum |
| quorum | — | The majority of nodes needed for a cluster to make decisions. Hence the odd node count |
| RAIDZ2 | — | ZFS double-parity mode: tolerates the loss of 2 disks per vdev. Space-efficient, slower than mirroring |
| RDIMM | Registered DIMM | Server memory module with a register, mandatory on EPYC |
| RPO | Recovery Point Objective | The maximum amount of data one accepts losing during a recovery |
| SAML | Security Assertion Markup Language | Legacy federated authentication protocol (XML), an alternative to OIDC that was not chosen |
| SAS / SATA | Serial Attached SCSI / Serial ATA | Disk interfaces. SAS is the server variant (12 Gb, dual port) |
| SFP28 / SFP+ | Small Form-factor Pluggable | Network transceiver cages: SFP+ up to 10 Gb, SFP28 up to 25 Gb (backward compatible) |
| SIEM | Security Information and Event Management | Centralised collection and correlation of security events (Wazuh) |
| SOC | Security Operations Center | Security monitoring (intrusions, logs, compliance) |
| special vdev | — | A group of fast disks in a ZFS pool dedicated to metadata. Greatly accelerates PBS. Losing it destroys the pool: always mirror it |
| SPOF | Single Point Of Failure | A component whose failure alone interrupts the service |
| SSO | Single Sign-On | One account (edu-ID) for all services |
| TLS | Transport Layer Security | Connection encryption (the "S" in HTTPS) |
| U.2 | — | NVMe SSD in a 2.5-inch enclosure, connectable in a hot-swap bay |
| vdev | Virtual Device | A group of disks forming one redundancy unit in a ZFS pool. A pool is an assembly of vdevs |
| VLAN | Virtual LAN | Logical segmentation of a physical network into isolated zones |
| VM | Virtual Machine | — |
| VPN | Virtual Private Network | Encrypted tunnel giving access to the internal network from outside |
| ZFS | Zettabyte File System | Filesystem with built-in snapshots, compression, checksums and redundancy |