18.3 Remote Tables & Federation

1. What a remote table represents

A Datasphere remote table provides access to data in a connected source without necessarily copying the data into Datasphere.

SAP uses HANA virtual-table technology for remote access.

2. Runtime

Consumer
   |
Datasphere view
   |
remote table
   |
virtual table
   |
remote source

The query may cross the platform boundary.

3. Federation is a performance contract

A federated design must define:

• expected query frequency,
• source concurrency,
• predicate selectivity,
• join behavior,
• network path,
• acceptable latency,
• source protection.

4. Statistics

SAP Datasphere provides a Remote Queries monitor and supports creating statistics for remote tables to help the HANA SQL optimizer make better decisions.

Statistics are especially important when cardinality and distribution affect query planning.

5. Pushdown

Prefer execution close to the data when the source can efficiently perform filtering and aggregation.

Do not assume every complex semantic operation will push down.

Validate with actual execution evidence.

6. Federation versus replication

Dimension	Federation	Replication
Data freshness	Potentially immediate	Depends on replication
Source dependency	High	Lower during consumption
Local storage	Minimal	Required
Query predictability	Source-dependent	More predictable
Source protection	Harder	Easier
Recovery	Source-dependent	Local copy available

7. Troubleshooting

If a federated query is slow:

1. identify the remote table,
2. inspect Remote Queries,
3. inspect generated execution behavior,
4. check statistics,
5. reduce unnecessary columns,
6. improve predicates,
7. assess join order,
8. consider replication.

Architect exercise

Design a federated customer lookup where 20,000 users may query simultaneously.

Determine whether federation is safe. Define a fallback persistence strategy.