HANA slow, crashing, or mid-migration.Get a verified HANA architect in minutes.

Yes. PM scopes the slowest statements in the first 15 minutes using the SQL Plan Cache and HANA Studio Performance Trace. The architect identifies the root causes that typically surface post-ECC migration: row store tables that should be column store, missing or stale optimizer statistics, calculation views written as nested SQLScript loops instead of set-based graphical models, and unpartitioned tables exceeding the 2-billion-record limit. The fix usually involves rewriting the worst calculation views to push filters down to the lowest data layer, adding partitioning (range or hash) on large fact tables, converting procedural SQLScript to table functions, and enabling the unfolding/instantiation that lets the HANA engine prune columns. The architect re-runs Explain Plan and Visualize Plan to confirm the optimizer chooses the column engine, then documents before-and-after runtimes so you can verify the improvement.

PM scopes the migration path in the first session based on your current revision and data volume. The architect assesses whether you can use the SAP HANA Cloud Central self-service migration, the Cloud Migration Tool, or a backup-and-restore with HSR (HANA System Replication) for near-zero downtime. Key decisions include sizing the HANA Cloud instance using the Capacity Units model rather than fixed memory tiers, configuring Native Storage Extension (NSE) to keep warm data on disk and cut memory cost, mapping on-premise SLT or SDI replication flows to Cloud-compatible connectivity (Cloud Connector, SDA), and validating that custom roles, XSA artifacts, and HDI containers carry over. The architect produces a runbook with cutover steps, rollback criteria, and a data-validation checklist so business teams sign off before go-live.

Yes. The architect designs the model around HANA strengths: column-store tables for analytical fact data, calculation views (dimension, cube with star join) for the semantic layer, and table functions for reusable logic. They define the data foundation, set up proper key and foreign-key relationships, and choose between graphical and SQLScript views based on whether the workload is set-based aggregation (graphical) or complex procedural transformation (SQLScript). The architect applies modeling best practices: pushing joins and filters to the lowest node, avoiding row-by-row cursors, using input parameters and variables for runtime filtering, and exposing the model through HDI containers with calculation views consumable by SAC, Power BI, or an OData service. They also plan data tiering with NSE and partitioning strategy up front so the platform scales past initial data volumes without redesign.

PM treats OOM as an incident and the architect diagnoses it in the first session. They review the HANA Memory Overview, the M_HEAP_MEMORY and M_CS_TABLES views, and the OOM trace files to identify whether the cause is oversized column tables loaded fully into memory, runaway intermediate result sets from poorly written calculation views, or a too-low global allocation limit. Fixes depend on the cause: enable Native Storage Extension to page cold data out of memory, partition and load-unload large tables on a schedule, add result-set size limits and statement memory limits per user/workload class, and rewrite the calculation views that materialize huge intermediate results. The architect also right-sizes the global_allocation_limit and configures workload classes so a single runaway report cannot starve the whole system. They confirm stability with a load test before closing.

Yes. The architect first identifies which runtime your landscape uses, because the approach differs significantly. Classic XS (XS Classic) is deprecated and SAP recommends migrating to XS Advanced (XSA) or HANA Cloud HDI. If you are on XS Classic, the architect can maintain existing .xsodata and .xsjs services while planning a migration to HDI containers, Node.js or Java microservices, and CDS-based data definitions. For XSA and HANA Cloud, they work natively with HDI containers, design-time artifacts (.hdbcalculationview, .hdbtable, .hdbrole), the Cloud Foundry deployment model, and grants files for cross-container access. The architect handles the .hdiconfig and .hdinamespace setup, build dependencies, and the synonym/grant pattern that lets one container securely read another container objects without exposing the whole schema.