New: prepared statements, UPDATE and DELETE, and key lookups that seek — 0.143 ms through the gateway.Read the update

Use case · Power & Utilities

Power & Utilities

Grid operators keep two kinds of data that rarely live well together: dense measurement streams and slow-changing asset records. UltraFabric stores both under one schema model and one verifier.

Workload shape

WritesHigh-rate measurements
ReadsRange scans and aggregates
TenancyRegions, subsidiaries, partners
ScaleBillions of points over time
Where the design fits

How the data lands

Design fit, not a customer claim: which parts of UltraFabric carry which parts of this workload.

  • Measurement streams as columnar collections, grouped into row groups with min and max zone maps
  • Assets, outages and work orders as rows with idempotent writes
  • Per-utility or per-region tenants with analytics across the operator's own data
In SQL

What it looks like

Only the feeder, at and kw columns are decoded for this plan; the rest of each row group is never read.

CREATE TABLE measurements (id bigint PRIMARY KEY, feeder text NOT NULL,
  at timestamp, kw numeric(10,2), kvar numeric(10,2)) WITH (storage = columnar);

SELECT feeder, max(kw), min(kw), avg(kw)
  FROM measurements WHERE at >= '2026-10-20' GROUP BY feeder;
Status

Built today, coming next

Everything in the first column is in the repository and tested; the second column is the roadmap in order.

Built today

  • Columnar scans read only the columns a query observes
  • Frame-of-reference and bit-packed encodings for numeric and time columns
  • Answers checked against DuckDB on every benchmark run

Next

  • Zone-map predicate skipping once the security-aware pushdown rule lands
  • A time-series engine measured on TSBS
  • Multi-node storage and replication

Tell us about yours.

Early access starts with a conversation about the workload you would move first.