aboutsummaryrefslogtreecommitdiffstats
path: root/docs
diff options
context:
space:
mode:
authorMaciek Konstantynowicz <mkonstan@cisco.com>2020-03-06 12:39:26 +0000
committerMaciek Konstantynowicz <mkonstan@cisco.com>2020-03-06 15:44:12 +0000
commit3f831590b3b71cd1f1d4a46e3b13ece832041f26 (patch)
tree992a3a84b5c52e5082a57568edb43e2c2db96403 /docs
parent18857b0c9cbc840f72df6e7e325d08821944200a (diff)
IETF: Update to PLRsearch draft.
Change-Id: Id1c9684c42c19eccfdc347e43221e097aed10514 Signed-off-by: Maciek Konstantynowicz <mkonstan@cisco.com>
Diffstat (limited to 'docs')
-rw-r--r--docs/ietf/draft-vpolak-bmwg-plrsearch-03.md (renamed from docs/ietf/draft-vpolak-bmwg-plrsearch-02.md)44
1 files changed, 30 insertions, 14 deletions
diff --git a/docs/ietf/draft-vpolak-bmwg-plrsearch-02.md b/docs/ietf/draft-vpolak-bmwg-plrsearch-03.md
index f725024d28..b09b5ce0be 100644
--- a/docs/ietf/draft-vpolak-bmwg-plrsearch-02.md
+++ b/docs/ietf/draft-vpolak-bmwg-plrsearch-03.md
@@ -1,8 +1,8 @@
---
title: Probabilistic Loss Ratio Search for Packet Throughput (PLRsearch)
# abbrev: PLRsearch
-docname: draft-vpolak-bmwg-plrsearch-02
-date: 2019-07-08
+docname: draft-vpolak-bmwg-plrsearch-03
+date: 2020-03-06
ipr: trust200902
area: ops
@@ -38,10 +38,16 @@ normative:
RFC8174:
informative:
+
+ FDio-CSIT-PLRsearch:
+ target: https://docs.fd.io/csit/rls2001/report/introduction/methodology_data_plane_throughput/methodology_plrsearch.html
+ title: "FD.io CSIT Test Methodology - PLRsearch"
+ date: 2020-02
+
draft-vpolak-mkonstan-bmwg-mlrsearch:
target: https://tools.ietf.org/html/draft-vpolak-mkonstan-bmwg-mlrsearch
title: "Multiple Loss Ratio Search for Packet Throughput (MLRsearch)"
- date: 2019-07
+ date: 2020-02
--- abstract
@@ -56,7 +62,8 @@ probabilistic in nature, and not deterministic.
# Motivation
-Network providers are interested in throughput a system can sustain.
+Network providers are interested in throughput a networking system can
+sustain.
[RFC2544] assumes loss ratio is given by a deterministic function of
offered load. But NFV software systems are not deterministic enough.
@@ -74,6 +81,10 @@ for benchmarking networking setups such as software based NFV systems.
# Terms And Assumptions
+Due to the indeterministic nature of certain NFV systems that are the
+targetted by PLRsearch algorithm, existing network benchmarking terms
+are explicated and a number of new terms and assumptions are introduced.
+
## Device Under Test
In software networking, "device" denotes a specific piece of software
@@ -647,7 +658,12 @@ direct critical load estimate, so users can draw their own conclusion.
Alternatively, trend analysis may be a part of exit conditions,
requiring longer searches for systems displaying trends.
-# Sample Implementation Specifics: FD.io CSIT
+# Known Implementations
+
+The only known working implementation of PLRsearch is in Linux
+Foundation FD.io CSIT open-source project [FDio-CSIT-PLRsearch].
+
+## FD.io CSIT Implementation Specifics
The search receives min_rate and max_rate values, to avoid measurements
at offered loads not supporeted by the traffic generator.
@@ -658,7 +674,7 @@ agnostic to flows and directions, it only cares about overall counts of
packets sent and packets lost), but debug output from traffic generator
lists unidirectional values.
-## Measurement Delay
+### Measurement Delay
In a sample implemenation in FD.io CSIT project, there is roughly 0.5
second delay between trials due to restrictons imposed by packet traffic
@@ -679,7 +695,7 @@ increasing trial durations with configurable coefficients (currently 5.1
seconds for the first trial, each subsequent trial being 0.1 second
longer).
-## Rounding Errors and Underflows
+### Rounding Errors and Underflows
In order to avoid them, the current implementation tracks natural
logarithm (instead of the original quantity) for any quantity which is
@@ -688,7 +704,7 @@ Python), so special value "None" is used instead. Specific functions for
frequent operations (such as "logarithm of sum of exponentials") are
defined to handle None correctly.
-## Fitting Functions
+### Fitting Functions
Current implementation uses two fitting functions. In general, their
estimates for critical rate differ, which adds a simple source of
@@ -735,7 +751,7 @@ compute. At the end, both fitting function implementations contain
multiple "if" branches, discontinuities are a possibility at range
boundaries.
-### Stretch Function
+#### Stretch Function
The original function (before applying logarithm) is Primitive Function
to Logistic Function. The name "stretch" is used for related a function
@@ -747,7 +763,7 @@ given as InputForm of Wolfram language:
r = (a*(1 + E^(m/a))*Log[(E^(b/a) + E^(m/a))/(1 + E^(m/a))])/E^(m/a)
-### Erf Function
+#### Erf Function
The original function is double Primitive Function to Gaussian Function.
The name "erf" comes from error function, the first primitive to
@@ -760,7 +776,7 @@ given as InputForm of Wolfram language:
r = ((a*(E^(-((b - m)^2/a^2)) - E^(-(m^2/a^2))))/Sqrt[Pi] + m*Erfc[m/a]
+ (b - m)*Erfc[(-b + m)/a])/(1 + Erf[m/a])
-## Prior Distributions
+### Prior Distributions
The numeric integrator expects all the parameters to be distributed
(independently and) uniformly on an interval (-1, 1).
@@ -776,7 +792,7 @@ The "stretch" parameter is generated simply as the "mrr" value raised to
a random power between zero and one; thus it follows a Reciprocal
Distribution.
-## Integrator
+### Integrator
After few measurements, the posterior distribution of fitting function
arguments gets quite concentrated into a small area. The integrator is
@@ -809,7 +825,7 @@ new area the posterior distribution is concentrated at. The second phase
(dominated by whole sample population) is actually relevant for the
critical rate estimation.
-## Offered Load Selection
+### Offered Load Selection
First two measurements are hardcoded to happen at the middle of rate interval
and at max_rate. Next two measurements follow MRR-like logic,
@@ -865,6 +881,6 @@ networks.
# Acknowledgements
-..
+To be added.
--- back \ No newline at end of file