Scale Computing SC//Platform (SC//HyperCore)

Single-node appliances are supported; a highly available cluster starts at three nodes.

Scale says clusters grow from a single edge node to eight nodes in the data centre; more capacity means more clusters under Fleet Manager.

Native single-node appliance configurations (HC1000, HC1150DF, HC5150D) with built-in replication to another node or cluster as often as every 5 minutes; no local HA.

No true 2-node HA cluster; options are a replicated pair of single nodes (async, manual failover) or two nodes plus a storage-only node acting as local witness.

No stretched or metro cluster.

Nodes of different sizes and storage types can be mixed; a storage-only mode exists for witness use, but no compute-only nodes.

Existing SANs can be used over iSCSI or NFS (LUNs/shares attached for VMs, backup and archive), outside SCRIBE protection.

HC3000 and HC5000 appliances are all-NVMe (U.3); HE100 uses M.2 NVMe; HEAT tiers flash and HDD in hybrid nodes.

Current Theory of Operations and data sheet describe space-efficient snapshots and clones (shared, reference-counted blocks) but no inline deduplication or compression of primary data; replication traffic is compressed.

Not available; SCRIBE always keeps two mirrored copies (RAID 10 equivalent).

Two copies of every chunk across nodes; tolerates a disk or a whole node failure, re-mirrors automatically into free space without hot spares. One failure domain level only.

Allocate-on-write VM snapshots (thousands per VM with little impact), instant promote/revert, instant thin clones and file-level recovery by mounting cloned disks.

No storage-layer encryption at rest; customers encrypt in the guest (BitLocker, LUKS) using a vTPM. Replication is AES-256 encrypted in flight.

No IOPS limits; HEAT gives per-virtual-disk flash priority (0-11 slider) on tiered nodes.

No native file services; partner options such as Nasuni run as VMs.

No native S3 object storage.

Not supported; SCRIBE storage serves only VMs on the cluster.

Built-in per-VM cluster-to-cluster replication, change-tracked and compressed, running continuously; Scale cites replication as often as every 5 minutes. Included at no extra cost.

Not available.

One-click promote of replicated snapshots and reverse replication for failback; runbooks and DR tests must be scripted with Ansible or the REST API.

Snapshots, file-level restore and remote replication are built in (Scale sells this as integrated backup/DR); no separate catalogued backup application.

Snapshots are not documented as immutable; immutable backups come via Veeam integration.

Veeam 12.3.2 supports SC//HyperCore (agentless integration); Acronis and Storware partners; fewer options than VMware.

No run-in-public-cloud recovery; Scale mentions a DR-as-a-service option and partners such as Wasabi and Acronis for cloud copies.

Standard VLAN-based networking with bonded LAN and backplane; no overlay SDN. SD-WAN is a separate product (SC//Connect).

No distributed firewall; traffic is directed only to specified virtual NICs and an internal firewall isolates the cluster.

Per-cluster RBAC (e.g. read-only, backup roles); OpenID Connect to Azure AD/ADFS with MFA enforced by the IdP; Fleet Manager SSO via Google or Microsoft.

Closed, vendor-controlled hypervisor stack with internal firewall and isolated backplane; UEFI Secure Boot and vTPM for VMs. No STIG baseline published.

Security practices aligned with ISO/IEC 27001 and SOC 2 (per Scale); appliances imaged in an ISO 9001 facility. No FIPS 140 or Common Criteria listing found.

SC//Fleet Manager manages 1 to 50,000 clusters: zero-touch provisioning, fleet updates, application lifecycle management, Secure Link remote access, Fleet Metrics and Fleet APIs.

One-click, rolling, non-disruptive updates of hypervisor, storage and firmware; minor updates apply in place without reboots; fleet-wide from Fleet Manager.

REST API, Terraform provider, Red Hat certified Ansible collection, cloud-init, plus fleet-level REST APIs in Fleet Manager.

AIME self-healing monitors hardware, software and VMs and remediates automatically; Fleet Metrics keeps up to a year of CPU, memory and storage history for planning.

No tenant self-service portal, quotas or chargeback.

Live migration, automatic VM restart within minutes after node failure, affinity rules and automatic placement; no DRS-style continuous balancing documented.

No integrated Kubernetes; run containers or K8s in VMs with a container-optimised OS. Fleet Manager ALM and partner Avassa help deploy containerised edge apps.

Management plane (Fleet Manager) is cloud-hosted, but there is no HyperCore instance in public clouds.

VDI partners are Parallels, Leostream and 10ZiG; no Citrix or Omnissa Horizon validation listed.

Edge AI positioning with GPU appliances (e.g. HC5000 with NVIDIA L4 24 GB) for inferencing and computer vision; no vGPU-sharing platform or AI stack.

No database service.

Nodes ship pre-imaged; rack, cable and initialise in minutes, or use Fleet Manager zero-touch provisioning (Scale claims 90% less install time).

Scale appliances (HE100, HE500, HC1000, HC3000, HC5000) or SC//HyperCore on HCL hardware via the imaging service, including Dell VxRail 14G-16G nodes.

Purpose-built for edge: NUC-size nodes with about 4 GB hypervisor overhead, self-healing, ZTP and fleet management of thousands of sites.

Clusters run fully locally with no management VM or cloud dependency; Fleet Manager is optional. Updates for disconnected sites need manual staging.

SC//Move (powered by Carbonite) for near-zero-downtime migration, Acronis convert-to-VM, Veeam restore, and Scale migration services; documented VxRail re-imaging path.

No free edition; a browser-based Remote Edition sandbox for prospects and a Workbench Edition for partners.

Same subscription model on Scale appliances or certified third-party servers; not usable in public clouds.

On Scale appliances ScaleCare covers hardware and software, including next-business-day part shipment in the US.

Severity 1 initial response objective 2 hours (24x7 by phone); Scale reports phone calls answered in 13 seconds on average (2023).

Smaller ecosystem: Lenovo, Supermicro, Intel, SNUC, Veeam, Acronis and niche edge partners; active user community but fewer skilled admins in the market.

Proprietary storage and management on a KVM base; runs on commodity certified hardware and VMs can be exported, but it is a single-hypervisor platform.