Single service comparison

Dataplex is a serverless control plane that federates governed access across GCS and BigQuery without owning data; its entities (lakes, zones, assets), tasks, and dataplex-managed IAM policies have no direct API equivalent in Huawei. DataArts Lake Formation provides a metastore plus centralized fine-grained authorization over OBS data and compatible engines (DLI, MRS, GaussDB(DWS)), while DataArts Fabric offers integrated governance logic in the DataArts Studio suite. The two Huawei services are not interchangeable: Lake Formation is the authorization/metastore layer; Fabric is the governance integration surface.

API and data-model scope differ materially. Dataplex exposes Lakes/Zones/Assets and Tasks APIs with native data-quality, lineage, and discovery built into the same product. Huawei splits these across DataArts Studio modules (Catalog, Governance, Quality, Lineage) and Lake Formation's RBAC/ABAC and metadata APIs, so parity must be validated feature-by-feature rather than assumed. Dataplex's automatic metadata extraction from GCS/BigQuery has no single Huawei counterpart; expect to compose Catalog jobs, Lake Formation metastore sync, and Quality modules to reproduce equivalent behavior.

Scaling, HA, and operational ownership diverge. Dataplex is fully managed and serverless, with Google handling scheduling and regional reliability; customers focus on policy and rule authoring. Huawei Lake Formation and DataArts Fabric are managed but rely on paired serverless/cluster compute engines (DLI serverless, MRS, DWS) for actual processing, and regional availability, quotas, and engine compatibility must be confirmed per region. Pricing also shifts from GCP processing-unit/dataproc-units metering to Huawei DLI/DWS/MRS runtime or cluster-spec billing plus OBS storage and cross-region transfer.

Start with an assessment that inventories Dataplex lakes, zones, assets, automated tasks, data-quality rules, lineage sources, and IAM policies. Decide target topology: DataArts Lake Formation for centralized metastore and fine-grained access, and DataArts Studio (Catalog, Quality, Lineage, Governance) hubbed through DataArts Fabric for cataloging and lifecycle governance. Do not assume one-to-one parity; map each Dataplex group/zone/asset to Huawei catalog entities and OBS prefixes, and confirm regional availability of Lake Formation, DLI, MRS, and DWS in each target region before designing cutover.

For data and configuration migration, move underlying object data to OBS using CDM or DRD where supported, then rebuild metadata in Lake Formation metastore and DataArts Catalog rather than mirroring Dataplex's lakes/zones model directly. Re-author data-quality rules in DataArts Quality and reconfigure lineage collection for DLI/MRS/DWS jobs; replace Dataplex IAM and IAM Conditions with Lake Formation RBAC/ABAC policies and IAM on OBS. Plan task reimplementations (Dataplex Tasks -> DataArts Factory jobs or DLI scheduled SQL) since Dataplex task APIs are not portable.

Validate governance behavior and cut over cautiously. Run parallel metadata, quality, and lineage checks against curated datasets and confirm policy enforcement on DLI/MRS/DWS queries before retiring Dataplex controls. Reconcile lineage completeness and access audits for a representative period. Cut over incrementally by domain or lake zone to limit blast radius, since neither Huawei provides an automated one-click Dataplex migration path.

Account for cost-model and operational changes in production. GCP bills Dataplex processing units and Dataproc serverless units plus storage; Huawei bills DLI/DWS/MRS runtime or cluster capacity plus OBS storage and intra/inter-region transfer, so recompute TCO using peak concurrency, rule frequency, scan volume, and retention. Also plan for quota requests (Lake Formation metastore, DLI compute, MRS clusters) and engine-version compatibility checks that have no Dataplex equivalent, and budget for any partner-supported professional services where native tooling gaps exist.