Mr.PlanB Logo
    Virtualization Storage Ledger 2026
    Ledger
    Virtualization Storage Ledger 2026

    Longhorn: Virtualization Storage Ledger 2026 Scorecard

    Replicated block storage for KubeVirt VMs inside SUSE Virtualization; the V2 engine is GA upstream but still a preview where it serves VMs.

    Edition October 16, 20268 min read

    Where Longhorn put its effort, 2021 to 2026

    Each square is one capability area in one year. The brighter it is, the more the area gained in the Virtualization Storage Ledger that year; 1, 2 and 3 mark that year's biggest pushes.

    ’21’22’23’24’25’26
    Data services core2
    Resilience and HA1
    Performance and protocols23
    Hypervisor integration11
    Scale and topology
    Data protection and DR2
    Security1
    Operations and automation3
    Hardware and deployment freedom3
    Kubernetes and containers
    no change+0.5+1+1.5 or moreScores out of 5 per area.

    Biggest pushes overall: Hypervisor integration (+2), Resilience and HA (+1), Performance and protocols (+1). Already strong in 2020: Kubernetes and containers, Hardware and deployment freedom, Resilience and HA.

    Top focus each year: 2021 Hypervisor integration; 2024 Security; 2025 Hypervisor integration; 2026 Resilience and HA. 2 of 6 years show no gain in any area.

    Part of the Virtualization Storage Ledger (edition 2026-10-16). How the scores work: methodology.

    Snapshot

    • Type: community project (CNCF Incubating), packaged by SUSE as SUSE Storage
    • Owner: The Longhorn Authors under the CNCF (Sandbox 2019-10-11, Incubating 2021-11-04); started by Rancher Labs, which SUSE bought in 2020; SUSE is private under EQT
    • Licence: Apache 2.0; free community builds; SUSE Storage and SUSE Virtualization sold through Rancher Prime subscriptions, prices not published
    • Latest release: 1.13.0 (2026-09-29); inside SUSE Virtualization: Harvester 1.9.0 (2026-09-16) with Longhorn 1.12.1
    • Support lifecycle: upstream: one year per release; SUSE Storage Prime: 1.12.1 EOM 2027-02-13 and EOL 2028-02-13; SUSE says production clusters should run Prime or Stable versions only
    • Next milestone: 1.12.2 (rc1 on 2026-10-15) and Harvester 1.9.1 with the fix for VM shutdown on upgrade; no dates published
    • Hypervisors: KubeVirt only, through SUSE Virtualization (Harvester) and plain KubeVirt on Kubernetes; no VMware, Hyper-V, Proxmox VE or XCP-ng (Longhorn README and Harvester release notes)
    • Tier: Kubernetes-native storage
    • Archetype: Dependable builder
    • Evidence grade: B. Release dates and contents for Longhorn 1.6, 1.8 and 1.10 to 1.13 and for Harvester 1.6 to 1.9 were read from GitHub, the SUSE lifecycle page and the Longhorn docs. Ownership history comes from Wikipedia. Longhorn 1.0 to 1.5, Harvester 1.1 to 1.5, partner terms and prices were not read, so the 2020 to 2023 steps and the 2023 direction record are thin.

    Scope of the scores: Longhorn is a Kubernetes CSI storage system. It is in this Ledger because it is the default VM storage of SUSE Virtualization (Harvester) and serves KubeVirt VM disks with synchronous replication, live migration and automatic failover. It cannot serve a classic hypervisor. It is scored as a community project with sustainability risk rows, because the open-source project is what SUSE packages and nobody else sells it.

    Scorecard

    DimensionScoreDirection
    Capability today60.3 / 100higher is better
    Momentum since end of 2022+14.3 points (2020 to 2022: +3.6)higher is better
    Credibility17.5 / 25higher is better
    Sustainability risk8.5 / 25lower is better
    Technical lock-in11.0 / 25lower is better
    Integrator fit12.0 / 20higher is better
    AI leverage2.0 / 20higher is better
    Ledger Index57.9 / 100balanced weights; see the methodology

    Capability by domain, 2015 to today

    Scores 0 to 5 against fixed anchors; total is weighted to 100.

    Domain (weight)201520192022Today
    Data services core (15)n/an/a22.5
    Resilience and HA (15)n/an/a2.53.5
    Performance and protocols (12)n/an/a1.52.5
    Hypervisor integration (12)n/an/a2.53
    Scale and topology (10)n/an/a22.5
    Data protection and DR (10)n/an/a2.53
    Security (8)n/an/a1.52.5
    Operations and automation (8)n/an/a2.53.5
    Hardware and deployment freedom (5)n/an/a3.54
    Kubernetes and containers (5)n/an/a44.5
    Total / 100n/an/a46.060.3

    Year by year: 2020: 42.4 · 2021: 46.0 · 2022: 46.0 · 2023: 46.0 · 2024: 49.3 · 2025: 52.8 · 2026: 60.3

    What moved the score

    YearDomainChangePointsTrigger
    2021Hypervisor integration1 → 2.5+3.6Harvester 1.0.0 GA (21 December; year inferred from the bundled Longhorn 1.2.3): default VM storage with live migration on KubeVirt
    2024Hardware and deployment freedom3.5 → 4+0.51.6.0: Talos and OKD support, V2 engine on ARM64 in preview; V1 ARM64 date not read
    2024Performance and protocols1.5 → 2+1.21.6.0: V2 data engine on SPDK in preview (half step)
    2024Security1.5 → 2.5+1.61.6.0: block volume encryption; Harvester 1.6.0 (2025) marks V1 volume encryption GA in SUSE Virtualization; no in-flight encryption or external key manager
    2025Data protection and DR2.5 → 3+1.01.8.0: several backup stores with a default target; Harvester 1.8.0 (2026) backs up guest-cluster PVCs
    2025Hypervisor integration2.5 → 3+1.2Harvester 1.6.0 and 1.7.0: V1 volume encryption GA, online resize from the VM page, RWX block volumes shared by several VMs, third-party multipath
    2025Kubernetes and containers4 → 4.5+0.51.8.0 and 1.10.0: automatic RWX expansion and CSIStorageCapacity; 1.11.0 (2026) adds ReadWriteOncePod and allowedTopologies
    2025Operations and automation2.5 → 3+0.81.8.0: Helm controller installs on RKE2 and K3s, automatic RWX volume expansion
    2026Data services core2 → 2.5+1.51.12.0 and 1.13.0: V2 data engine GA with fast linked clones and volume group snapshots; erasure-coded sharding still experimental
    2026Operations and automation3 → 3.5+0.81.11.0 and 1.13.0: S.M.A.R.T. disk health monitoring, V2 live upgrade from 1.12.2, global manager; no Terraform or Ansible provider
    2026Performance and protocols2 → 2.5+1.21.12.0: V2 data engine GA upstream; Harvester 1.9.0 ships it as Technical Preview, so the second half step waits
    2026Resilience and HA2.5 → 3.5+3.01.11.0 and 1.12.0: rebuild pulls from several replicas, balance-aware disk scheduling, topology-aware provisioning; no stretched or metro mode
    2026Scale and topology2 → 2.5+1.01.11.0 and 1.13.0: balance-aware disk scheduling and longhorn-global-manager for several clusters

    Direction, 2023 to 2026

    Each release item carries one theme and one driver (V vision, C customer need, M market window, U upstream, P portfolio).

    DateItemThemeDriver
    2024-021.6.0: block volume encryption, V2 preview gains snapshots, backup and ARM64, Talos and OKD supportSecurityC
    2025-011.8.0: several backup stores with a default target, automatic RWX expansion, Helm controller installs, V2 experimental live migrationData protectionC
    2025-08Harvester 1.6.0: third-party storage for guest clusters, V1 volume encryption GA, online expansion, V2 storage networkSecurityC
    2025-091.10.0: V2 engine interrupt mode, volume and snapshot cloning, replica rebuild QoS, volume expansion, IPv6 for V1PerformanceV
    2025-12Harvester 1.7.0: online resize from the VM page, V2 volume expansion, RWX block volumes to several VMs, third-party multipathHypervisor integrationC
    2026-011.11.0: rebuild from several replicas, balance-aware disk scheduling, S.M.A.R.T. monitoring, V2 engine to Technical Preview, hotfix imagesResilienceC
    2026-04Harvester 1.8.0: VM disks move between Longhorn and external CSI without downtime, backend-enforced Rancher rolesHypervisor integrationC
    2026-061.12.0: V2 data engine GA, topology-aware provisioning, IPv6 and dual stack for V2, V2 backing images removedPerformanceV
    2026-09Harvester 1.9.0: V2 engine to Technical Preview in Harvester, unified backup framework, concurrent live migration; notes advise against upgrading yetData protectionC
    2026-091.13.0: V2 live upgrade from 1.12.2, volume group snapshots, experimental erasure-coded sharding, global manager, Kubernetes 1.34 or laterOperationsC
    2026-101.12.2-rc1 and Harvester 1.9.1-rc3 with the permanent fix for VM shutdown on upgradeOperationsC
    • Centre of gravity: Security (2), Data protection (2), Performance (2)
    • Blind spots (no items): Data services, Scale, Hardware, Kubernetes, Packaging, Licensing, AI
    • Driver mix: Vision 18%, Customer need 82%, Market window 0%, Upstream 0%, Portfolio 0%
    • Idea-to-need index: 0.18 (0 = demand-pulled, 1 = thesis-pushed)

    Credibility: 17.5 / 25

    CriterionScoreBasis
    Cadence4Minors in 2025-09, 2026-01, 2026-06 and 2026-09, about every four to five months, with patch lines; no published calendar
    Roadmap transparency3.5Public repository, issue tracker and release notes that name planned items; no roadmap page
    Say-do3.5V2 engine went from Technical Preview (1.11, 2026-01) to GA (1.12, 2026-06) in four months; fast V2 cloning slipped from 1.12.1 to 1.13; Harvester still ships V2 as Technical Preview
    Velocity4Several substantive items per minor since 1.10
    Lifecycle stability2.5One year of upstream support per release; Kubernetes minimum jumped to 1.34 in 1.13; hotfix images for 1.10.0 and 1.11.0; Harvester 1.9.0 notes advise against upgrading

    Sustainability risk: 8.5 / 25 (lower is better)

    CriterionScoreBasis
    Maintainer concentration3SUSE engineers run the releases and the README routes support bundles and vulnerability reports to SUSE addresses; 595 contributing organisations in the CNCF record
    Backer health2Backed by SUSE, private under EQT since 2023; CNCF Incubating since 2021-11-04; contributors down 5 percent year on year
    Licence-change risk1Apache 2.0 under CNCF ownership, which blocks a unilateral relicence
    Abandonment signs0.5Four minors in twelve months and 1.12.2-rc1 on 2026-10-15
    Security response2Vulnerability reports go to SUSE; 1.13 gives the CSI service account cluster-wide read on Secrets by default; no public advisory process read

    Technical lock-in: 11.0 / 25 (lower is better)

    CriterionScoreBasis
    Formats1.5Replica directories managed by Longhorn; VM images enter as qcow2 or raw through CDI
    Export path2Backups to S3 or NFSv4; Harvester 1.8.0 moves VM disks to external CSI storage without downtime
    Stack coupling3.5Requires Kubernetes 1.34 or later for 1.13 and a KubeVirt platform for VMs; no classic hypervisor
    Hardware / cloud coupling1x86_64 or ARM64; Harvester recommends YES-certified hardware; ARM64 V2 has a known NVMe stuck IO
    Skills and tooling coupling3Kubernetes, Rancher and KubeVirt skills; longhornctl

    Integrator fit: 12.0 / 20

    CriterionScoreBasis
    Partner programme openness3Open-source download; SUSE channel terms not read
    Multi-tenancy and self-service3Namespaces, Kubernetes quotas and Rancher RBAC; Harvester enforces roles in the backend since 1.8
    API and infrastructure-as-code3.5CRDs, Helm and longhornctl; a Harvester Terraform provider was not verified
    Skills and certification2.5SUSE and Rancher training exists; Longhorn-specific certification not found

    AI leverage: 2.0 / 20

    How the product uses AI, scored per year like capability. Only GA counts.

    Criterion20192022TodayLatest step
    AI workload infrastructuren/a002020: None
    AI platform stackn/a002020: None
    AI-assisted operationsn/a002020: None
    Agent opennessn/a222020: Kubernetes CRDs and the Longhorn API; no MCP server

    Year by year: 2020: 2 · 2021: 2 · 2022: 2 · 2023: 2 · 2024: 2 · 2025: 2 · 2026: 2

    AI say-do, 2023 to 2026: 0 AI items announced; 0 GA, 0 still preview, 0 pending, 0 dropped. No AI item has been announced for Longhorn; the vGPU and MIG features SUSE lists belong to SUSE Virtualization.

    Longhorn has announced no AI feature. Agents can drive it through its CRDs like any Kubernetes resource, and the GPU features SUSE markets sit in SUSE Virtualization, not in the storage.

    Reading

    Longhorn scores as a solid Kubernetes volume manager that became VM storage because SUSE built a hypervisor on it. The data services and scale lines stay low because it has no compression, deduplication, QoS or two-node mode, and the performance line is held back half a step because the V2 engine that is GA upstream since 1.12.0 ships as a Technical Preview in Harvester 1.9.0. Resilience and operations have improved most since 2025, with multi-replica rebuilds, topology-aware placement and V2 live upgrades. The one thing a buyer should check is which Longhorn version their Harvester release pins and whether that release is a Prime or Stable version SUSE supports in production. Sustainability risk is modest because the CNCF owns the project, but the people who ship it work for SUSE.

    Open questions and items to re-score

    • Whether the V2 data engine counts as GA for VM use: GA upstream since 1.12.0 (2026-06-02), Technical Preview in Harvester 1.9.0 and 'in preview' on the SUSE Storage page.
    • Harvester 1.0.0 is dated 21 December on GitHub without a year; 2021 is inferred from the bundled Longhorn 1.2.3. Harvester 1.1 to 1.5 and Longhorn 1.0 to 1.5 were not read.
    • The SUSE lifecycle page has inconsistent dates (SUSE Storage 1.12.1 GA listed 2 Jun 2026 while GitHub shows 2026-08-14).
    • Whether OpenShift Virtualization or Spectro Cloud VMO support Longhorn for VMs was not checked.
    • Harvester 1.9.1 with the permanent fix for VM shutdown on upgrade was still a release candidate on 2026-10-08.
    • SUSE Storage and SUSE Virtualization prices and partner terms were not found.

    Related reading

    More from Virtualization Storage Ledger 2026