Single service comparison

Azure Managed Disks expose a single managed-disk abstraction across multiple performance tiers (Standard HDD, Standard SSD, Premium SSD, Ultra Disk) with per-disk bursting and, for Premium/Ultra, configurable IOPS and throughput independent of capacity. Huawei EVS is the analogous multi-tenant block service with tiers (General Purpose SSD, General Purpose SSD V2, Ultra-high I/O, Extreme SSD) and General Purpose SSD V2 also decouples IOPS/throughput from capacity. Direct parity varies by tier; Ultra Disk's sub-millisecond latency and fine-grained performance tuning have no exact EVS equivalent and require validation against Extreme SSD.

For workloads needing physically dedicated storage, Huawei offers DSS as a separate dedicated storage-pool service (High I/O and Ultra-high I/O pools, billed yearly, with raw-capacity tiers and linear scaling). Azure has no direct dedicated-pool abstraction; Ultra Disk and Premium SSD with ZRS provide the closest performance/isolation options. DSS requires committing to a whole storage pool and interconnecting it with ECS/BMS/DCC, shifting operational responsibility from per-disk provisioning to pool planning—an operating-model change, not a drop-in replacement.

Azure integrates disks natively with Azure Backup, snapshot, encryption at rest with platform or customer-managed keys (via Key Vault), and zone-redundant storage options. Huawei EVS snapshots, CBR backups, KMS-based encryption, and shared-disk attach (up to 16 servers) cover analogous concerns, but regional availability, ZRS-equivalent synchronous replication, and CMK behaviors differ and must be checked per region. Cross-region replication for disks is not a managed one-click feature on either platform; both require backup/restore or server-level replication tooling.

Start with assessment: inventory every Azure disk's sku tier, capacity, configured IOPS/throughput, bursting usage, encryption (platform/CMK), ZRS/LRS setting, snapshot count, and attached VM. Map each disk to the closest EVS tier (Premium SSD -> Ultra-high I/O or Extreme SSD; Ultra Disk -> Extreme SSD with performance validation; Standard SSD -> General Purpose SSD V2; Standard HDD -> General Purpose SSD). Reserve DSS only for workloads that explicitly need physically dedicated pools or yearly-commitment economics. Confirm target AZ and region support before committing.

For data migration, there is no automated first-party Azure-to-Huawei disk conversion path. Use a lift-and-shift pattern: provision equivalent EVS disks on the target ECS/BMS, then transfer data via block-level copy (snapshot export to object storage such as OBS then import, or application-level replication like rsync/LVM mirroring over a VPN/Direct Connect link). Azure snapshots export to VHD; rehydrate to an ECS instance and copy block data onto the new EVS disk. Capture VM OS disk and data disks separately, and reinitialize disk signatures/UUIDs to avoid boot conflicts on cutover.

Validate before cutover: run IOPS/throughput benchmarks (fio/crystal-like) against the new EVS disk and compare against the Azure baseline, accounting for capacity-independent provisioning on GPSSD V2 / Extreme SSD. Confirm encryption, KMS key mapping to Huawei KMS, snapshot/backup policies via CBR, shared-disk multi-attach (up to 16 servers) where WSFC/RAC is used, and monitoring via Cloud Eye. Plan rollback: keep the Azure disk and a known-good snapshot until validation completes.

Mind the cost and gaps. Azure bills per-disk capacity plus tier and, for Ultra, provisioned IOPS/throughput; EVS bills per GiB-month by tier with GPSSD V2 charging for provisioned IOPS/throughput, and DSS bills yearly per pool. Recalculate TCO including snapshot storage, backup retention, and cross-region/interconnect traffic—Huawei has no native disk-level ZRS-equivalent sync replication, so durability beyond a single AZ requires CBR or server-level replication. Confirm quota limits (disk size up to 32 TiB, IOPS/throughput caps per tier) in the target region before production.