Fixed cost per operation
Fixed cost per operation implies a natural batch size.
What it is
Every medium has a natural transfer size. Divide the fixed cost by the transfer rate and you find the size at which half of your time is setup. Below it you mostly pay for setup.
From Physical Substrate & Performance Laws › The dial and the fixed cost
A fixed cost per operation produces a natural batch size, and the fixed cost moves upward. Every operation has a setup cost F paid no matter how big the operation is. If you handle B records per operation, the per-record cost is F ÷ B. Inside one machine, F was a disk seek, then a system call. Here F is the network round trip, the broker request, the API call. The formula does not change. Only the dominant F changes. Batching is the same mechanism, but it now has a semantic cost too: a batch fails as a unit.
From Data Movement › The epistemic limit
Where it appears
In 20 mechanisms. Each says what the mechanism is about, then how this principle shows up there, in its own words.
- Physical Substrate & Performance Laws
Why do trade-offs exist at all?
This is why chatty protocols fail over long distances, and why combining requests is the universal fix.
Read it inStorage media: what each one is made of and what it forcesThe network: a shared medium with invisible neighbors
- Data Movement
How do records cross between two systems that have independent rates, independent availability and independent lifetimes, and what are you willing to lose when the assumption underneath the crossing breaks?
Batch while the downstream is busy; flush immediately when it is idle.
Read it inBatchingWhat changed: the epistemic limitPartitioning: the four strategies
- Storage: Physical Organization
What should sit next to what, given that I must decide at write time and find out at read time?
This is the batch-size rule: when each operation has a fixed cost, the natural unit of work is the fixed cost divided by the cost per byte.
Read it inThe cost of one accessObject storage and the lakehouseData at rest and the one ratioMutation strategies: what happens when a value changesWhat physical organization forbidsChoose a mutation strategy
- Persistence & Durability
When a system says "yes, your data is safe", what exactly has it promised, and what did it have to do to keep that promise?
More important: the fixed cost moves upward.
Read it inGroup commitWrite-ahead logging (redo)Battery-backed memory and NVRAM
- Core Data Structures
How do we choose a data structure when the real cost is set by the hardware, not by Big-O?
And as hardware gets faster, the fixed cost does not vanish; it moves up into software.
Read it inThe big picture: a data structure is a betImmutable and persistent data structuresB-trees and B+treesSorted runs and the sparse indexSorted runs and merge structures
- Indexing
When does a separate "where is it?" structure pay for itself, and what does it cost to keep it true?
This is the same rule that gives the natural transfer size of every storage medium: the fixed cost of an access determines the granularity of the summary above it.
Read it inComposite indexes: everything follows from lexicographic orderDensity: sparse indexes and zone mapsHigh dimension: exactness becomes unaffordable
- Query Planning & Optimization
How does a system choose a plan for a question when the facts that decide the best plan are unknown until the plan has run?
Staleness is the result of that amortization, not an oversight.
Read it inStatisticsCompile or interpret, materialize or pipelineSpeed up a reasonable plan
- Caching
What happens when we deliberately keep a second copy of data that no protocol forces to be correct?
Second, and this is worth keeping: P/S is a natural batch size, with reuse count as the batch.
Read it inWrite routing, and why you delete rather than updateThe governing arithmetic: two inequalitiesThe amortization inequality is the \"natural batch size\" in a new costumeChoose what to cacheSize a cache
- Distributed Systems Coordination
Who is allowed to decide, and how does everyone else learn the decision without asking?
The precondition is what makes this a mechanism instead of a habit: in-band state must be interpretable with no shared context.
Read it inIn-band coordinationPartitioning and the partition count
- Fault Tolerance & Delivery Semantics
What is the cheapest thing that can be safely done twice?
This is the same batching arithmetic as before (pay a fixed cost less often), with the duplicate count as the thing being traded.
Read it inConsumer offsets and commit orderingWrite-ahead intent and the recovery scanWhy failure recovery is a different kind of problemChoose a recovery mechanism
- Scheduling & Workload Management
Which work runs now, which work waits, and who pays for the waiting?
> Locality scheduling does not track disks.
Read it inDelay schedulingLocality: node, rack, zone, anyCost-aware placement and spot
- Cost Mechanics
What is metered, and what is the smallest increment you are charged for?
A minimum increment is the fixed cost per operation of the vendor, shown on your bill.
Read it inLifecycle, tiering and retentionMaterialization break-evenCompute: capacity-time, not workPer-request pricing: the envelope and the letter
- Change Data Capture & Incremental Processing
What does it cost to know what changed, and is that cheaper than not needing to know?
At small scale incremental is strictly worse: you pay the complexity and receive nothing.
Read it inThe arithmetic that governs everything
- Table & Storage Maintenance
Who pays for the gap between the order data arrives in and the order it is read in: the writer, every reader, or a background job, and what is the break-even?
total cost per unit of time = C / T + r · T / 2, which is smallest at T = √(2C / r).
Read it inOptimal file size: it is F divided by bDelete compaction and orphan detectionTrigger policy: when should maintenance fire?Verification of the rewrite
- Data Quality & Validation
How do you check that data is right, when "right" is a claim about the world and not about the bytes?
One fact dominates: type checking is free because decoding is not optional.
Read it inThe reconciliation ladder: detecting is cheap, locating is what costsSchema validation and the per-value tier
- Metadata, Catalog & Lineage
What can a system prove about its data without reading the data, and how does it know that proof is still true?
Write throughput has a hard ceiling of one successful commit per catalog round trip, however much data each commit carries.
Read it inThe commit pointStatistics stores
- Orchestration & Workflow
How do you make unreliable, unobservable, irreversible side effects behave like a pure function of a time interval?
The mechanism is correct when the work per unit is much larger than the launch cost, and wrong when it is not.
Read it inDynamic task generation
- Security & Privacy
How do you protect data from someone who watches your design and then chooses how to attack it?
The fixed cost of one key-service round trip sets the natural batch size: one data key per file, per partition or per table.
- Specialized Processing Models
What do you build when a query predicate cannot be reduced to a range over an ordered key?
Visibility latency equals the flush interval: a document is unsearchable until it lands in a segment.
Read it inSegments and near-real-time refresh
- Measurement & Experimental Method
What would have to be true for this number to mean what I think it means?
Sampling costs by clock, instrumentation costs by event; one finds the bulk, the other finds the instance.
Read it inSampling, instrumentation and hardware countersPublishing configs and one-command rerunsMeasurement overhead and wall versus CPU time
Related
Made on the same pages most often.