A time-series database, because each order has a creation timestamp and orders accumulate continuously over time as the business operates.

Any database type performs equivalently here, because the workload is simple and the differences between database types do not apply.

A relational database, because every order has structured fields that map cleanly to columns in a single orders table.

A document database, because each order is a self-contained record with nested fields that are read and written together

The latency between the candidate region and the existing user base, since user-perceived performance depends on the round-trip time.

The cost difference between the candidate region and the existing region, since cloud pricing varies between regions.

The data residency rules in the candidate jurisdiction, since some jurisdictions cannot legally hold the application's records.

The carbon intensity of the candidate region's electricity grid, since cleaner grids reduce the application's emissions.

Edge nodes located near user populations hold cached copies of content and serve most requests without contacting the origin.

The CDN encrypts traffic between the origin and the user, which removes the encryption work from the origin server itself.

Edge nodes apply load balancing across multiple origin servers so that no single origin server has to handle the full traffic volume.

The CDN compresses content before sending it to users, which reduces the size of each response and the work the origin must do to serve it.

The application has been configured to update the cache before every read, which doubles the work performed on each request.

The cache uses too much memory on the application servers and is being flushed to disk before it can serve any requests.

The cache hit rate is too low, so most requests miss the cache and continue through to the database with little load reduction.

The database is faster than the cache for the data being requested, which makes the cache layer counterproductive in this case.

The replication stream from the primary needs the secondary's network equipment to be active even when no servers are running.

Regulatory rules in most jurisdictions require that a backup site demonstrate continuous operational readiness through power consumption.

The standby servers must remain powered and synchronized so that the secondary region can take over quickly when the primary fails.

Cloud providers bill for the secondary region whether or not its servers are actively powered on at any given moment.

A flow battery, because a four-to-ten-hour discharge window makes the technology the most adaptable option across any grid service.

A lithium-ion battery sized for twelve hours of discharge, because longer discharge duration always takes priority over faster response.

A flywheel system, because very fast response within seconds is the exact requirement and flywheels are built for this short-duration, high-cycle duty.

A pumped-hydro reservoir, because the large energy capacity ensures that any frequency event is covered regardless of duration.

None of the three, because a 2-megawatt-hour battery is too small to be useful in any of the three roles described for this warehouse.

Option Z, because backup power is the traditional purpose of grid-scale batteries and should be prioritized regardless of the actual outage risk at the site.

Option X, because demand response events carry environmental value and that sustainability benefit should outweigh any purely financial consideration in this decision.

Option Y, because recurring monthly savings across all twelve months deliver the largest ongoing financial benefit at a site with few outages.

Smart-city platforms require field-level millisecond response times and no available cloud infrastructure is capable of meeting that latency requirement.

Several organizations with different data formats, privacy rules, and commercial incentives were involved at once, and governance agreement has not been reached.

The hub-and-spoke architecture commonly used in smart cities is technically inferior to a fully centralized design and should never have been selected.

BACnet, Modbus, and KNX are incompatible at the protocol level and cannot be bridged across a city-scale platform under any technical conditions.

Curtailing the refrigeration compressors completely for the full duration of the event in order to maximize the size of the reduction.

Installing rooftop solar on every store in the chain so that each store becomes self-powered during summer afternoon events.

Shifting the entire store operation into overnight hours so that all consumption falls outside the demand response event window.

Cycling refrigeration and easing back-of-house setpoints while leaving the shop floor conditions and customer-facing areas unchanged.

Manual response, because a human operator can apply judgement and produce more consistent reductions than any automated demand response system.

Full BMS replacement, because only a complete rebuild of the building automation stack can guarantee acceptable long-term performance across all subsystems.

No participation, because a BACnet-era BMS cannot be integrated with any current demand response programme under existing standards.

A gateway that translates OpenADR to BACnet, because the existing BMS investment is preserved while automated response is added at a low capital cost.