OTAP Parquet study - #3405

Closed

jmacd wants to merge 18 commits into

open-telemetry:mainfrom

jmacd:jmacd/parquet_study

Commits on Jul 1, 2026

chore(benchmarks): add otap_parquet study of OTAP/IPC vs flattened Parquet

jmacd
and
Copilot
committed
chore(benchmarks): add Vortex contender to the otap_parquet study

jmacd
and
Copilot
committed
chore(benchmarks): study Vortex write throughput at 10k/100k block sizes

jmacd
and
Copilot
committed
chore(benchmarks): drop Vortex and add per-step pipeline breakdown

jmacd
and
Copilot
committed
docs(benchmarks): add ANALYSIS.md explaining the measured cost ratios

jmacd
and
Copilot
committed
feat(benchmarks): account for OTAP/IPC streaming amortization

jmacd
and
Copilot
committed
docs(benchmarks): split the streaming savings into schema and dictionaries

jmacd
and
Copilot
committed
docs(benchmarks): explain flat IPC steady-state (no cross-frame compression)

jmacd
and
Copilot
committed
docs(benchmarks): quantify the cross-batch redundancy IPC leaves unexploited

jmacd
and
Copilot
committed
docs(benchmarks): add gateway-precompute perspective to the analysis

jmacd
and
Copilot
committed
docs(benchmarks): correct the .NET IPC compression facts in the analysis

jmacd
and
Copilot
committed
docs(benchmarks): trim the .NET IPC compression tangent from the analysis

jmacd
and
Copilot
committed
docs(benchmarks): keep the analysis logs-specific

jmacd
and
Copilot
committed
feat(benchmarks): study OTAP-flat single columnar view and pipeline conversions

jmacd
and
Copilot
committed
docs(benchmarks): explain the direct standard-to-Parquet path and its ordering precondition

jmacd
and
Copilot
committed
feat(benchmarks): measure naive vs optimized OTAP-to-Parquet flatten

jmacd
and
Copilot
committed
docs(benchmarks): make the parquet study vendor-neutral

jmacd
and
Copilot
committed
docs(otap_parquet): pair and tighten the two analysis docs

jmacd
and
Copilot
committed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

OTAP Parquet study - #3405

OTAP Parquet study#3405
jmacd wants to merge 18 commits into
open-telemetry:mainfrom
jmacd:jmacd/parquet_study

Commits on Jul 1, 2026

chore(benchmarks): add otap_parquet study of OTAP/IPC vs flattened Parquet

chore(benchmarks): add Vortex contender to the otap_parquet study

chore(benchmarks): study Vortex write throughput at 10k/100k block sizes

chore(benchmarks): drop Vortex and add per-step pipeline breakdown

docs(benchmarks): add ANALYSIS.md explaining the measured cost ratios

feat(benchmarks): account for OTAP/IPC streaming amortization

docs(benchmarks): split the streaming savings into schema and dictionaries

docs(benchmarks): explain flat IPC steady-state (no cross-frame compression)

docs(benchmarks): quantify the cross-batch redundancy IPC leaves unexploited

docs(benchmarks): add gateway-precompute perspective to the analysis

docs(benchmarks): correct the .NET IPC compression facts in the analysis

docs(benchmarks): trim the .NET IPC compression tangent from the analysis

docs(benchmarks): keep the analysis logs-specific

feat(benchmarks): study OTAP-flat single columnar view and pipeline conversions

docs(benchmarks): explain the direct standard-to-Parquet path and its ordering precondition

feat(benchmarks): measure naive vs optimized OTAP-to-Parquet flatten

docs(benchmarks): make the parquet study vendor-neutral

docs(otap_parquet): pair and tighten the two analysis docs