mcssimulation
3 years ago- apparently, using the sigma (σ) sign causes font-related PDF/A validation errors... pdfa
spio [Wed, 10 Aug 2016 08:09:45 +0000 (10:09 +0200)]
- apparently, using the sigma (σ) sign causes font-related PDF/A validation errors, while the mu (µ) sign is allowed: CIDSet, CIDToGIDMap, CMAPs

4 years ago- introduction of generic properties which are simply captured as uninterpreted strin... master bomba/master
Josef Spillner [Sun, 4 Oct 2015 07:53:42 +0000 (09:53 +0200)]
- introduction of generic properties which are simply captured as uninterpreted string values
- this will enable a plugin-based overall characteristic prediction such as overall expected response time depending on the number of fragments per service

4 years ago- graphical handling of optimisation preferences in addition to the optimisation...
Josef Spillner [Sun, 4 Oct 2015 07:37:49 +0000 (09:37 +0200)]
- graphical handling of optimisation preferences in addition to the optimisation constraints

4 years ago- preliminary implementation of context menu, still without real action origin/HEAD origin/master
Josef Spillner [Fri, 25 Sep 2015 13:48:13 +0000 (15:48 +0200)]
- preliminary implementation of context menu, still without real action

4 years ago- improved highlighted display of the significant and redundant fragments
Josef Spillner [Fri, 25 Sep 2015 13:41:38 +0000 (15:41 +0200)]
- improved highlighted display of the significant and redundant fragments

4 years ago- more generic configuration, controllable with environment variables
Josef Spillner [Thu, 20 Aug 2015 08:57:29 +0000 (10:57 +0200)]
- more generic configuration, controllable with environment variables

4 years ago- basic support for network proxy availability emulation, currently hardcoded for...
Josef Spillner [Mon, 6 Jul 2015 11:33:45 +0000 (13:33 +0200)]
- basic support for network proxy availability emulation, currently hardcoded for destination port :80

4 years ago- extended introductory documentation with a mention of the emulation options
Josef Spillner [Wed, 1 Jul 2015 14:48:39 +0000 (16:48 +0200)]
- extended introductory documentation with a mention of the emulation options

4 years ago- avoid output when CSV is activated
Josef Spillner [Wed, 1 Jul 2015 14:13:24 +0000 (16:13 +0200)]
- avoid output when CSV is activated
- more reasonable default availabilities: 97% for all containers

4 years ago- more flexible handling of different start/stop execution backend
Josef Spillner [Wed, 1 Jul 2015 12:58:51 +0000 (14:58 +0200)]
- more flexible handling of different start/stop execution backend
- implementation of a remote execution of docker commands, interpreting container instances as services
- CNA scenario file for the docker backend

4 years ago- improved unavailability logic of the incident model
Josef Spillner [Wed, 1 Jul 2015 08:42:13 +0000 (10:42 +0200)]
- improved unavailability logic of the incident model
- log of about one hour of dynamic simulation

4 years ago- introduction of distinct unavailability models: convergence (as before) and incident
Josef Spillner [Wed, 1 Jul 2015 07:17:05 +0000 (09:17 +0200)]
- introduction of distinct unavailability models: convergence (as before) and incident
- minimal single-service simulation script with a single 90% availability service
- preparation of plotfile to compare unavailability models graphically

4 years ago- capturing the single-service emulation traces to demonstrate unavailability patterns
Josef Spillner [Fri, 5 Jun 2015 05:35:05 +0000 (07:35 +0200)]
- capturing the single-service emulation traces to demonstrate unavailability patterns
- new variant 'extralong' to see how convergence improves over time; however, this will need another uninterrupted run (e.g. in a VM)

4 years ago- flavour of the experiment for very long-running emulation (10k seconds)
Josef Spillner [Wed, 3 Jun 2015 16:54:25 +0000 (18:54 +0200)]
- flavour of the experiment for very long-running emulation (10k seconds)
- attention: random is slightly incomplete
- will require even longer runs to see absolutely clear convergence

4 years ago- plot files for all measurements so far
Josef Spillner [Tue, 2 Jun 2015 21:44:28 +0000 (23:44 +0200)]
- plot files for all measurements so far

4 years ago- complete measurement on the goldstandard.ini dataset with 90% target availability
Josef Spillner [Tue, 2 Jun 2015 20:54:23 +0000 (22:54 +0200)]
- complete measurement on the goldstandard.ini dataset with 90% target availability

4 years ago- bugfix: dynamic k-threshold is actually valid for all cases, the number of candidat...
Josef Spillner [Tue, 2 Jun 2015 14:14:11 +0000 (16:14 +0200)]
- bugfix: dynamic k-threshold is actually valid for all cases, the number of candidate services does just not matter

4 years ago- renewed measurements: about 1000 data points, totalling 2500 seconds, for each...
Josef Spillner [Tue, 2 Jun 2015 13:25:33 +0000 (15:25 +0200)]
- renewed measurements: about 1000 data points, totalling 2500 seconds, for each algorithm

4 years ago- first results from emulating scenario.ini for 1000+ seconds with the permissible...
Josef Spillner [Tue, 2 Jun 2015 11:54:12 +0000 (13:54 +0200)]
- first results from emulating scenario.ini for 1000+ seconds with the permissible methods for 60% minimum availability

4 years ago- handling of replication special case
Josef Spillner [Tue, 2 Jun 2015 11:53:57 +0000 (13:53 +0200)]
- handling of replication special case

4 years ago- basic CSV export support
Josef Spillner [Tue, 2 Jun 2015 09:17:45 +0000 (11:17 +0200)]
- basic CSV export support

4 years ago- new wrapper module which selects the right algorithm module [only supports availabi...
Josef Spillner [Tue, 2 Jun 2015 08:53:59 +0000 (10:53 +0200)]
- new wrapper module which selects the right algorithm module [only supports availability for now]
- integration of proper distribution determination into the emulation client by using the wrapper module

4 years ago- new 'stable' option which hamstrings on/off switches
Josef Spillner [Tue, 2 Jun 2015 08:12:13 +0000 (10:12 +0200)]
- new 'stable' option which hamstrings on/off switches

4 years ago- corresponding client part with random retrievals to check the actual availability...
Josef Spillner [Sun, 24 May 2015 10:54:50 +0000 (12:54 +0200)]
- corresponding client part with random retrievals to check the actual availability of a file
- currently, this supports only k=3/m=0 equal fragment distribution without redundancy

4 years ago- first support for independent emulation of service availabilities with local FTP...
Josef Spillner [Sun, 24 May 2015 09:44:30 +0000 (11:44 +0200)]
- first support for independent emulation of service availabilities with local FTP/HTTP servers

4 years ago- improved aesthetical diagrams
Josef Spillner [Tue, 21 Apr 2015 22:28:24 +0000 (00:28 +0200)]
- improved aesthetical diagrams

4 years ago- improved/consolidated 3D plots
Josef Spillner [Tue, 21 Apr 2015 09:38:04 +0000 (11:38 +0200)]
- improved/consolidated 3D plots

4 years ago- PICav+ sweetspot plot
Josef Spillner [Sat, 18 Apr 2015 20:38:25 +0000 (22:38 +0200)]
- PICav+ sweetspot plot

4 years ago- sweetspot-related description file
Josef Spillner [Sat, 18 Apr 2015 20:35:37 +0000 (22:35 +0200)]
- sweetspot-related description file
- some annotations for improving the PICav(+) implementation found while formalising the algorithm

4 years ago- improved PICav+ implementation which is now competitive with a sweetspot of 8 servi...
Josef Spillner [Fri, 17 Apr 2015 20:39:24 +0000 (22:39 +0200)]
- improved PICav+ implementation which is now competitive with a sweetspot of 8 services (on rumba)

4 years ago- first integration of PICav code into PICav+
Josef Spillner [Fri, 17 Apr 2015 17:54:07 +0000 (19:54 +0200)]
- first integration of PICav code into PICav+

4 years ago- explicit configurability of approximated availability calculations inside each...
Josef Spillner [Fri, 17 Apr 2015 16:51:42 +0000 (18:51 +0200)]
- explicit configurability of approximated availability calculations inside each module
- two more scenario descriptions of 15 and 20 services, respectively

4 years ago- removal of target parameters from non-guaranteeing fixedproportional module
Josef Spillner [Thu, 16 Apr 2015 20:55:27 +0000 (22:55 +0200)]
- removal of target parameters from non-guaranteeing fixedproportional module
- addition of new pseudo-calculation module for random targets selection and assignment

4 years ago- start with PICav+ implementation using Staggered as baseline
Josef Spillner [Thu, 16 Apr 2015 20:17:29 +0000 (22:17 +0200)]
- start with PICav+ implementation using Staggered as baseline

4 years ago- new calculation mode fixed:replication
Josef Spillner [Thu, 16 Apr 2015 18:46:17 +0000 (20:46 +0200)]
- new calculation mode fixed:replication

4 years ago- adaptation of the graphical frontend to the latest library code
Josef Spillner [Thu, 16 Apr 2015 18:33:22 +0000 (20:33 +0200)]
- adaptation of the graphical frontend to the latest library code

4 years ago- improved display of services
Josef Spillner [Thu, 16 Apr 2015 18:23:19 +0000 (20:23 +0200)]
- improved display of services
- work with equal availabilities or prices

4 years ago- introduction of a new calculation mode "absolute" which assigns all data to one...
Josef Spillner [Thu, 16 Apr 2015 18:12:32 +0000 (20:12 +0200)]
- introduction of a new calculation mode "absolute" which assigns all data to one particular preferred sink
- more explicit assignment which now excludes certain services by setting fragment=0

4 years ago- sync of experiment with corrected Monte Carlo code
Josef Spillner [Thu, 16 Apr 2015 13:26:08 +0000 (15:26 +0200)]
- sync of experiment with corrected Monte Carlo code

4 years ago- cleanups and corrections in the Monte Carlo code
Josef Spillner [Thu, 16 Apr 2015 13:15:24 +0000 (15:15 +0200)]
- cleanups and corrections in the Monte Carlo code

4 years ago- complementary plots for n=10, without much differences
Josef Spillner [Thu, 16 Apr 2015 07:01:12 +0000 (09:01 +0200)]
- complementary plots for n=10, without much differences

4 years ago- first empirical experiments on the quality of Monte Carlo approximations
Josef Spillner [Thu, 16 Apr 2015 06:37:37 +0000 (08:37 +0200)]
- first empirical experiments on the quality of Monte Carlo approximations

4 years ago- fixed approximation function
Josef Spillner [Wed, 15 Apr 2015 21:58:52 +0000 (23:58 +0200)]
- fixed approximation function
- transparent integration of this function as alternative mode into the interfaces

4 years ago- first implementation of the Monte Carlo approximation for estimating the resulting...
Josef Spillner [Wed, 15 Apr 2015 21:16:42 +0000 (23:16 +0200)]
- first implementation of the Monte Carlo approximation for estimating the resulting overall availability from a distribution scheme

4 years ago- adaptive display depending on the number of services in the calc-availability script
Josef Spillner [Wed, 15 Apr 2015 20:16:56 +0000 (22:16 +0200)]
- adaptive display depending on the number of services in the calc-availability script
- new simulation of ten services with equal capacity and price but decreasing availability

4 years ago- simple introduction docs
Josef Spillner [Wed, 15 Apr 2015 19:17:54 +0000 (21:17 +0200)]
- simple introduction docs

4 years ago- introduction of top-level output colouring to make the results easily distinguishable
Josef Spillner [Wed, 15 Apr 2015 05:56:50 +0000 (07:56 +0200)]
- introduction of top-level output colouring to make the results easily distinguishable

4 years ago- increase ε to 10ms to account for sometimes rather rough scheduling of calculation...
Josef Spillner [Wed, 15 Apr 2015 05:49:28 +0000 (07:49 +0200)]
- increase ε to 10ms to account for sometimes rather rough scheduling of calculation steps

4 years ago- make staggered calculation honour time limits as well
Josef Spillner [Wed, 15 Apr 2015 05:48:03 +0000 (07:48 +0200)]
- make staggered calculation honour time limits as well

4 years ago- additional a-posteriori runtime checking within a 2ms epsilon boundary
Josef Spillner [Wed, 15 Apr 2015 05:34:01 +0000 (07:34 +0200)]
- additional a-posteriori runtime checking within a 2ms epsilon boundary

4 years ago- price checking and logging in the combinatoric distribution search
Josef Spillner [Wed, 15 Apr 2015 05:24:22 +0000 (07:24 +0200)]
- price checking and logging in the combinatoric distribution search
- a-posteriori price checking in the calculation script

4 years ago- correction of propagation of cumulated price from calculator models to the top...
Josef Spillner [Tue, 14 Apr 2015 21:48:06 +0000 (23:48 +0200)]
- correction of propagation of cumulated price from calculator models to the top-level script
- first basic handling of maximum prices in the staggered module

4 years ago- completion of the calling interface with max-price and max-runtime parameters
Josef Spillner [Tue, 14 Apr 2015 06:28:50 +0000 (08:28 +0200)]
- completion of the calling interface with max-price and max-runtime parameters
- max-price is already handled to staggered, but not handled yet
- max-runtime is handled to combinatory and handled correctly

4 years ago- fusion between combinatoric and staggered: Now staggered delivers a staggeredcombin...
Josef Spillner [Tue, 14 Apr 2015 05:25:28 +0000 (07:25 +0200)]
- fusion between combinatoric and staggered: Now staggered delivers a staggeredcombinatoric method which is on par but capacity-aware

4 years ago- bugfix/workaround: division by zero
Josef Spillner [Mon, 13 Apr 2015 05:13:49 +0000 (07:13 +0200)]
- bugfix/workaround: division by zero

4 years ago- improved staggering for small quantities
Josef Spillner [Mon, 13 Apr 2015 05:09:32 +0000 (07:09 +0200)]
- improved staggering for small quantities

4 years ago- adaptive debug depending on whether a comparison between all distribution schemes...
Josef Spillner [Mon, 13 Apr 2015 04:59:46 +0000 (06:59 +0200)]
- adaptive debug depending on whether a comparison between all distribution schemes is wanted or not

4 years ago- vague coarse-grained redundancy setting over all slices in the staggered distributi...
Josef Spillner [Sun, 12 Apr 2015 21:45:35 +0000 (23:45 +0200)]
- vague coarse-grained redundancy setting over all slices in the staggered distribution method

4 years ago- integration of the staggered distribution determination into the experimentation...
Josef Spillner [Sun, 12 Apr 2015 21:20:22 +0000 (23:20 +0200)]
- integration of the staggered distribution determination into the experimentation and simulation tools

4 years ago- optimised plot of the staggered distribution
Josef Spillner [Sun, 12 Apr 2015 20:54:28 +0000 (22:54 +0200)]
- optimised plot of the staggered distribution

4 years ago- new distribution calculation module for staggered distribution slices
Josef Spillner [Sun, 12 Apr 2015 20:03:01 +0000 (22:03 +0200)]
- new distribution calculation module for staggered distribution slices
- experimentation script for parameter combinations with appropriate plotting file

4 years ago- script clean-up to prepare it for module inclusion
Josef Spillner [Sun, 12 Apr 2015 19:41:13 +0000 (21:41 +0200)]
- script clean-up to prepare it for module inclusion

4 years ago- first code towards the calculation of staggered redundancies, availabilities and...
Josef Spillner [Sun, 12 Apr 2015 19:02:22 +0000 (21:02 +0200)]
- first code towards the calculation of staggered redundancies, availabilities and capacities

4 years ago- foreseeable integration of PICav+
Josef Spillner [Sun, 12 Apr 2015 08:34:42 +0000 (10:34 +0200)]
- foreseeable integration of PICav+

4 years ago- correction of proportional distribution implementation: account for real proportions
Josef Spillner [Sun, 12 Apr 2015 08:09:51 +0000 (10:09 +0200)]
- correction of proportional distribution implementation: account for real proportions

4 years ago- first really basic module to host the PICav+ algorithm for efficient multi-criteria...
Josef Spillner [Tue, 7 Apr 2015 21:22:26 +0000 (23:22 +0200)]
- first really basic module to host the PICav+ algorithm for efficient multi-criteria optimisation

4 years ago- first very basic support for multi-criteria proportional assignments of m-fragments...
Josef Spillner [Tue, 7 Apr 2015 21:10:27 +0000 (23:10 +0200)]
- first very basic support for multi-criteria proportional assignments of m-fragments; k-fragments yet to do

4 years ago- ignore bytecode files
Josef Spillner [Tue, 7 Apr 2015 12:01:20 +0000 (14:01 +0200)]
- ignore bytecode files

4 years ago- bugfix: the modules may change the redundancy level so it needs to be reset in...
Josef Spillner [Tue, 7 Apr 2015 10:02:49 +0000 (12:02 +0200)]
- bugfix: the modules may change the redundancy level so it needs to be reset in this script before each module invocation

4 years ago- documentation bugfix
Josef Spillner [Mon, 6 Apr 2015 21:14:45 +0000 (23:14 +0200)]
- documentation bugfix

5 years ago- initial import of the MCS optimisation simulation, including the MCS-SIM graphical...
Josef Spillner [Thu, 8 Jan 2015 18:07:42 +0000 (19:07 +0100)]
- initial import of the MCS optimisation simulation, including the MCS-SIM graphical simulator and the corresponding command-line simulators