Thanks a lot for your comments, Mark. Since our new Student Information System will be open source we would be more than happy to make contributions back into the community. Since our SIS will be based on SOA principles, we are thinking of Making Drools a decision service (web service) for a single entry point since we don't want to have thousands of rule services. Thanks, Len ________________________________ From: rules-users-bounces(a)lists.jboss.org [mailto:rules-users-bounces@lists.jboss.org] On Behalf Of Mark Proctor Sent: Saturday, November 17, 2007 9:44 AM To: Rules Users List Subject: Re: [rules-users] Drools and BRMS (60,000+ rules) Carlsen, Len wrote: Hello, I am wondering if anyone has any experience in managing very large rule sets and facts and could share their experiences in the design and management of this. We are embarking on a new Student Information System project here at UBC in conjunction with several other American universities. We have pretty much decided to use Drools as our rules engine and maybe also the Drools BRMS for our business rules management. Our first application is the course curriculum application/module where we will manage course co-requisites, pre-requisites, etc. and we estimate that we will have about 20,000 course rules. We currently have about 15,000 course rules (QuickRules) in production now. We will also have security rules, input/form validation rules, student awards rules, degree rules, enrolment rules, admission rules, student financial rules etc. which will total about 40,000 rules. So currently we are looking at managing about 60,000 rules (not including versioning). If we include versioning then there would probably be more than 100,000 rules. We are using jackrabbit JCR, I think it can handle this, but I don't have any hard figures for the scalability. If it works for you, please do let us know :) Example: A single course could have 10 or more rules per course version. We currently have about 2000 course and most courses have more than 1 version. We are looking into developing our own rules management system to define dependencies between rules, courses (facts) and course + rule versions. E.g. Course CHEM101 version 1 links to CHEM101 a single rule or a rule set/package version 1 by a rule id. We need the ability to notify users when rules are modified which impacts other rules and what impact changing a rule can cause for example a degree program. E.g. changing a course's pre-requisite may affect other courses' pre-requisites; rules about rules. We also need the ability to find out why a student was not able to register for a course etc. We will also need to show the course dependency graph visually (the RDBMS can probably help us here). Later, we are hoping to use the Drools Solver for course and exam scheduling and also to use it for student degree planning and student awards. We don't currently have an exposed event model in the BRMS or change notification. However this would not be hard to add, maybe you could help us? We have drools-analytics which will create an api with the change impact requirements you need, it is not yet currently integrated into the BRMS, bit if you do the notificatino part it should be easy enough to add. We will need to translate the rules into English language syntax for publication in the course calendar and later into other languages (first French) since the SIS application will be used internationally. Listing requirements could just be shown in bulleted form. Probably can't use the DSL in this case since we would need to support several languages or maybe I am wrong here? DSL are just a templating language, if can be made to look english like - if that's what you need. You also have the "docementation" field in the BRMS, which you can use. If you have other ideas please spec them out in a JIRA and we'll look inoto doing them in the future, or maybe you can help us :) If we use the Drools BRMS to store all of our rules on a relational database, will we be able to write SQL queries to get at specific rules/packages to link facts (courses) to rules? Or do we have to go through the BRMS to get at the rules. Does BRMS have a facility to validate rules against facts so you can see if your rules execute correctly against your data/facts before committing your rules. Or maybe we need a testing framework for this. We use JackRabbit JCR, we provide high level apis to accessing data, but you an ofcourse ue the jackrabbit low level apsi - if you really know what you are doing. We have the ability to validate the compilation of the rules, scenario testing is currently being added and will be in the next release. Other issues are the performance of the Drools Solver and having thousands of rules compiled, loaded and running. During registration we could have thousands of rules executing concurrently. The rules engine and the BRMS would be services as part of an SOA infrastructure (and an OSGi service environment). Anyone got any experience with Drools in an OSGi environment with respect to class loading issues when compiling rules and facts from other bundles since rules and facts may be in different bundles? Bundle buddy class loading? No but it is something that interests us, please give us your OSGi feedback. Solver is not currently a supported product, its very much R&D alphaware, but I'm sure geoffrey would love your feedback, good to see people are evaluating it. Later, we will need translators to/from other rule engines; maybe ruleML and JCR can help us here. JCR is jus a storage repository, for rule translators you are on your own. However Drools itself is language independant, we are exposing the api for pluggeable parsers - so in theory you could write parsers for each of these. So far, I think Drools can do most of this; am I correct here? Does any of this sound plausible? I'm sure you can get by for now, and work with us to make it the best system for your needs. Thanks very much for any tips, thoughts or comments. Len Len Carlsen Enrolment Services - Student Systems University of British Columbia ________________________________ _______________________________________________ rules-users mailing list rules-users(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users

We did some simple micro performance/scalability testing a while back on Drools 4.0.1 where we came to a similar conclusion. We may even publish our test results (and improve our tests) in the future. Len ________________________________ From: rules-users-bounces(a)lists.jboss.org [mailto:rules-users-bounces@lists.jboss.org] On Behalf Of Mark Proctor Sent: Tuesday, November 20, 2007 12:46 PM To: Rules Users List Subject: Re: [rules-users] Drools and BRMS (60,000+ rules) Carlsen, Len wrote: Thanks a lot for your comments, Mark. Since our new Student Information System will be open source we would be more than happy to make contributions back into the community. Since our SIS will be based on SOA principles, we are thinking of Making Drools a decision service (web service) for a single entry point since we don't want to have thousands of rule services. Thanks, Len Mic's been doing some scalability profiling for you. Engine execution can scale to 60K rules, however the compilation and build process is quite slow at that large numbers. The good news is we have never profiled and optimised our compilation and build processes, we haven't had the use cases yet to warrant this - so we can definitely make some improvements here, when we get time :) ________________________________ From: rules-users-bounces(a)lists.jboss.org [mailto:rules-users-bounces@lists.jboss.org] On Behalf Of Mark Proctor Sent: Saturday, November 17, 2007 9:44 AM To: Rules Users List Subject: Re: [rules-users] Drools and BRMS (60,000+ rules) Carlsen, Len wrote: Hello, I am wondering if anyone has any experience in managing very large rule sets and facts and could share their experiences in the design and management of this. We are embarking on a new Student Information System project here at UBC in conjunction with several other American universities. We have pretty much decided to use Drools as our rules engine and maybe also the Drools BRMS for our business rules management. Our first application is the course curriculum application/module where we will manage course co-requisites, pre-requisites, etc. and we estimate that we will have about 20,000 course rules. We currently have about 15,000 course rules (QuickRules) in production now. We will also have security rules, input/form validation rules, student awards rules, degree rules, enrolment rules, admission rules, student financial rules etc. which will total about 40,000 rules. So currently we are looking at managing about 60,000 rules (not including versioning). If we include versioning then there would probably be more than 100,000 rules. We are using jackrabbit JCR, I think it can handle this, but I don't have any hard figures for the scalability. If it works for you, please do let us know :) Example: A single course could have 10 or more rules per course version. We currently have about 2000 course and most courses have more than 1 version. We are looking into developing our own rules management system to define dependencies between rules, courses (facts) and course + rule versions. E.g. Course CHEM101 version 1 links to CHEM101 a single rule or a rule set/package version 1 by a rule id. We need the ability to notify users when rules are modified which impacts other rules and what impact changing a rule can cause for example a degree program. E.g. changing a course's pre-requisite may affect other courses' pre-requisites; rules about rules. We also need the ability to find out why a student was not able to register for a course etc. We will also need to show the course dependency graph visually (the RDBMS can probably help us here). Later, we are hoping to use the Drools Solver for course and exam scheduling and also to use it for student degree planning and student awards. We don't currently have an exposed event model in the BRMS or change notification. However this would not be hard to add, maybe you could help us? We have drools-analytics which will create an api with the change impact requirements you need, it is not yet currently integrated into the BRMS, bit if you do the notificatino part it should be easy enough to add. We will need to translate the rules into English language syntax for publication in the course calendar and later into other languages (first French) since the SIS application will be used internationally. Listing requirements could just be shown in bulleted form. Probably can't use the DSL in this case since we would need to support several languages or maybe I am wrong here? DSL are just a templating language, if can be made to look english like - if that's what you need. You also have the "docementation" field in the BRMS, which you can use. If you have other ideas please spec them out in a JIRA and we'll look inoto doing them in the future, or maybe you can help us :) If we use the Drools BRMS to store all of our rules on a relational database, will we be able to write SQL queries to get at specific rules/packages to link facts (courses) to rules? Or do we have to go through the BRMS to get at the rules. Does BRMS have a facility to validate rules against facts so you can see if your rules execute correctly against your data/facts before committing your rules. Or maybe we need a testing framework for this. We use JackRabbit JCR, we provide high level apis to accessing data, but you an ofcourse ue the jackrabbit low level apsi - if you really know what you are doing. We have the ability to validate the compilation of the rules, scenario testing is currently being added and will be in the next release. Other issues are the performance of the Drools Solver and having thousands of rules compiled, loaded and running. During registration we could have thousands of rules executing concurrently. The rules engine and the BRMS would be services as part of an SOA infrastructure (and an OSGi service environment). Anyone got any experience with Drools in an OSGi environment with respect to class loading issues when compiling rules and facts from other bundles since rules and facts may be in different bundles? Bundle buddy class loading? No but it is something that interests us, please give us your OSGi feedback. Solver is not currently a supported product, its very much R&D alphaware, but I'm sure geoffrey would love your feedback, good to see people are evaluating it. Later, we will need translators to/from other rule engines; maybe ruleML and JCR can help us here. JCR is jus a storage repository, for rule translators you are on your own. However Drools itself is language independant, we are exposing the api for pluggeable parsers - so in theory you could write parsers for each of these. So far, I think Drools can do most of this; am I correct here? Does any of this sound plausible? I'm sure you can get by for now, and work with us to make it the best system for your needs. Thanks very much for any tips, thoughts or comments. Len Len Carlsen Enrolment Services - Student Systems University of British Columbia ________________________________ _______________________________________________ rules-users mailing list rules-users(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users ________________________________ _______________________________________________ rules-users mailing list rules-users(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users

If we do publish them then we will re-run the tests against the latest release. Maybe we can send you the tests for validation if we decide to publish them. Len ________________________________ From: rules-users-bounces(a)lists.jboss.org [mailto:rules-users-bounces@lists.jboss.org] On Behalf Of Mark Proctor Sent: Tuesday, November 20, 2007 3:24 PM To: Rules Users List Subject: Re: [rules-users] Drools and BRMS (60,000+ rules) Carlsen, Len wrote: We did some simple micro performance/scalability testing a while back on Drools 4.0.1 where we came to a similar conclusion. We may even publish our test results (and improve our tests) in the future. sure, but if you could give us a bit to tweak trunk, so if we publish old stats we can say, but atleast its improving. We'll be focusing on this end of the year, beginning of the new year. Len ________________________________ From: rules-users-bounces(a)lists.jboss.org [mailto:rules-users-bounces@lists.jboss.org] On Behalf Of Mark Proctor Sent: Tuesday, November 20, 2007 12:46 PM To: Rules Users List Subject: Re: [rules-users] Drools and BRMS (60,000+ rules) Carlsen, Len wrote: Thanks a lot for your comments, Mark. Since our new Student Information System will be open source we would be more than happy to make contributions back into the community. Since our SIS will be based on SOA principles, we are thinking of Making Drools a decision service (web service) for a single entry point since we don't want to have thousands of rule services. Thanks, Len Mic's been doing some scalability profiling for you. Engine execution can scale to 60K rules, however the compilation and build process is quite slow at that large numbers. The good news is we have never profiled and optimised our compilation and build processes, we haven't had the use cases yet to warrant this - so we can definitely make some improvements here, when we get time :) ________________________________ From: rules-users-bounces(a)lists.jboss.org [mailto:rules-users-bounces@lists.jboss.org] On Behalf Of Mark Proctor Sent: Saturday, November 17, 2007 9:44 AM To: Rules Users List Subject: Re: [rules-users] Drools and BRMS (60,000+ rules) Carlsen, Len wrote: Hello, I am wondering if anyone has any experience in managing very large rule sets and facts and could share their experiences in the design and management of this. We are embarking on a new Student Information System project here at UBC in conjunction with several other American universities. We have pretty much decided to use Drools as our rules engine and maybe also the Drools BRMS for our business rules management. Our first application is the course curriculum application/module where we will manage course co-requisites, pre-requisites, etc. and we estimate that we will have about 20,000 course rules. We currently have about 15,000 course rules (QuickRules) in production now. We will also have security rules, input/form validation rules, student awards rules, degree rules, enrolment rules, admission rules, student financial rules etc. which will total about 40,000 rules. So currently we are looking at managing about 60,000 rules (not including versioning). If we include versioning then there would probably be more than 100,000 rules. We are using jackrabbit JCR, I think it can handle this, but I don't have any hard figures for the scalability. If it works for you, please do let us know :) Example: A single course could have 10 or more rules per course version. We currently have about 2000 course and most courses have more than 1 version. We are looking into developing our own rules management system to define dependencies between rules, courses (facts) and course + rule versions. E.g. Course CHEM101 version 1 links to CHEM101 a single rule or a rule set/package version 1 by a rule id. We need the ability to notify users when rules are modified which impacts other rules and what impact changing a rule can cause for example a degree program. E.g. changing a course's pre-requisite may affect other courses' pre-requisites; rules about rules. We also need the ability to find out why a student was not able to register for a course etc. We will also need to show the course dependency graph visually (the RDBMS can probably help us here). Later, we are hoping to use the Drools Solver for course and exam scheduling and also to use it for student degree planning and student awards. We don't currently have an exposed event model in the BRMS or change notification. However this would not be hard to add, maybe you could help us? We have drools-analytics which will create an api with the change impact requirements you need, it is not yet currently integrated into the BRMS, bit if you do the notificatino part it should be easy enough to add. We will need to translate the rules into English language syntax for publication in the course calendar and later into other languages (first French) since the SIS application will be used internationally. Listing requirements could just be shown in bulleted form. Probably can't use the DSL in this case since we would need to support several languages or maybe I am wrong here? DSL are just a templating language, if can be made to look english like - if that's what you need. You also have the "docementation" field in the BRMS, which you can use. If you have other ideas please spec them out in a JIRA and we'll look inoto doing them in the future, or maybe you can help us :) If we use the Drools BRMS to store all of our rules on a relational database, will we be able to write SQL queries to get at specific rules/packages to link facts (courses) to rules? Or do we have to go through the BRMS to get at the rules. Does BRMS have a facility to validate rules against facts so you can see if your rules execute correctly against your data/facts before committing your rules. Or maybe we need a testing framework for this. We use JackRabbit JCR, we provide high level apis to accessing data, but you an ofcourse ue the jackrabbit low level apsi - if you really know what you are doing. We have the ability to validate the compilation of the rules, scenario testing is currently being added and will be in the next release. Other issues are the performance of the Drools Solver and having thousands of rules compiled, loaded and running. During registration we could have thousands of rules executing concurrently. The rules engine and the BRMS would be services as part of an SOA infrastructure (and an OSGi service environment). Anyone got any experience with Drools in an OSGi environment with respect to class loading issues when compiling rules and facts from other bundles since rules and facts may be in different bundles? Bundle buddy class loading? No but it is something that interests us, please give us your OSGi feedback. Solver is not currently a supported product, its very much R&D alphaware, but I'm sure geoffrey would love your feedback, good to see people are evaluating it. Later, we will need translators to/from other rule engines; maybe ruleML and JCR can help us here. JCR is jus a storage repository, for rule translators you are on your own. However Drools itself is language independant, we are exposing the api for pluggeable parsers - so in theory you could write parsers for each of these. So far, I think Drools can do most of this; am I correct here? Does any of this sound plausible? I'm sure you can get by for now, and work with us to make it the best system for your needs. Thanks very much for any tips, thoughts or comments. Len Len Carlsen Enrolment Services - Student Systems University of British Columbia ________________________________ _______________________________________________ rules-users mailing list rules-users(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users ________________________________ _______________________________________________ rules-users mailing list rules-users(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users ________________________________ _______________________________________________ rules-users mailing list rules-users(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users

This is very cool and is exactly what we will need for our new open source SIS. Once we have designed our BRMS then we will look more closely at Drools Solver (especially after the ITC competition). Currently, we are using Ad Astra for our course timetabling and exam scheduling and it would be good if we could replace that with Solver. Thanks for your interest, Len -----Original Message----- From: rules-users-bounces(a)lists.jboss.org [mailto:rules-users-bounces@lists.jboss.org] On Behalf Of Geoffrey De Smet Sent: Friday, November 23, 2007 5:09 AM To: rules-users(a)lists.jboss.org Subject: [rules-users] Re: Drools and BRMS (60,000+ rules) Mark Proctor wrote: infrastructure R&D yes, feedback on solver is very welcome :) The good news is, I am now working on an example called Examination, which is an implementation for the itc2007 examination competition: http://www.cs.qub.ac.uk/itc2007/examtrack/exam_track_index_files/exameva luation.htm It's based on real-world data. The "starter" dataset contains 7883 students, 607 exams, 54 periods and 7 rooms. There are 5 hard constraints and 7 soft constraints: http://www.cs.qub.ac.uk/itc2007/examtrack/exam_track_index_files/exameva luation.htm The bad news is, it's not performing yet, because I need to make a scalable selector implementation, which I am working on. There's an example called lesson schedule (which schedules courses), but that's just a toy example to learn drools-solver, at it just has 2 hard constraints. One piece of advice though: separate the solver part from the rest. Once the solver starts calculating, you 'll only want to have score-related rules in the rulebase. The solver can easily do a million fireAllRules per minute, so every performance loss is blown up. With kind regards, Geoffrey De Smet _______________________________________________ rules-users mailing list rules-users(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users

We would probably use the existing server side code and extent the current BRMS client UI, to satisfy our needs. By designing our BRMS I mostly meant designing the structure of all our rule domains, rules/packages and facts (courses/learning units). Len ________________________________ From: rules-users-bounces(a)lists.jboss.org [mailto:rules-users-bounces@lists.jboss.org] On Behalf Of Mark Proctor Sent: Saturday, November 24, 2007 1:02 AM To: Rules Users List Subject: Re: [rules-users] Re: Drools and BRMS (60,000+ rules) Carlsen, Len wrote: This is very cool and is exactly what we will need for our new open source SIS. Once we have designed our BRMS then we will look more closely at Drools Solver (especially after the ITC competition). Are you writting your own BRMS from scratch? Are there any particular issues with the BRMS at the moment that would warrant this? Maybe you could work on the BRMS we have now to extend it to satisfy your needs. Currently, we are using Ad Astra for our course timetabling and exam scheduling and it would be good if we could replace that with Solver. Thanks for your interest, Len -----Original Message----- From: rules-users-bounces(a)lists.jboss.org [mailto:rules-users-bounces@lists.jboss.org] On Behalf Of Geoffrey De Smet Sent: Friday, November 23, 2007 5:09 AM To: rules-users(a)lists.jboss.org Subject: [rules-users] Re: Drools and BRMS (60,000+ rules) Mark Proctor wrote: Carlsen, Len wrote: Other issues are the performance of the Drools Solver and having thousands of rules compiled, loaded and running. During registration we could have thousands of rules executing concurrently. The rules engine and the BRMS would be services as part of an SOA infrastructure (and an OSGi service environment). Anyone got any experience with Drools in an OSGi environment with respect to class loading issues when compiling rules and facts from other bundles since rules and facts may be in different bundles? Bundle buddy class loading? No but it is something that interests us, please give us your OSGi feedback. Solver is not currently a supported product, its very much R&D alphaware, but I'm sure geoffrey would love your feedback, good to see people are evaluating it. yes, feedback on solver is very welcome :) The good news is, I am now working on an example called Examination, which is an implementation for the itc2007 examination competition: http://www.cs.qub.ac.uk/itc2007/examtrack/exam_track_index_files/exameva luation.htm It's based on real-world data. The "starter" dataset contains 7883 students, 607 exams, 54 periods and 7 rooms. There are 5 hard constraints and 7 soft constraints: http://www.cs.qub.ac.uk/itc2007/examtrack/exam_track_index_files/exameva luation.htm The bad news is, it's not performing yet, because I need to make a scalable selector implementation, which I am working on. There's an example called lesson schedule (which schedules courses), but that's just a toy example to learn drools-solver, at it just has 2 hard constraints. One piece of advice though: separate the solver part from the rest. Once the solver starts calculating, you 'll only want to have score-related rules in the rulebase. The solver can easily do a million fireAllRules per minute, so every performance loss is blown up. With kind regards, Geoffrey De Smet _______________________________________________ rules-users mailing list rules-users(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users _______________________________________________ rules-users mailing list rules-users(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/rules-users

Hi Len. Yes the main challenges I see are to do with the sheer numbers of rules you see yourself happening, and how to manage that in any sort of BRMS. ie are there 100's of packages, or a few uber packages etc. How do you break it up so its manageable? etc... even simple things, like if there is a drop down with rule names - that doesnt' scale to well to that number of rules (in a single package). We have categories which allow you to break down views into rules etc. The other side of the coin is compile time taken and memory used. Good to see you have done some profiling of it (I saw you PDF) - I haven't had a chance to load up your project yet but I did do some of my own profiling. Since 4.0.3 the impact of deserialization and rule packages has been partly addressed. I am worried that one rule change would require a 5 minute rebuild of an entire package (is 5 minutes too long? I really don't know how much your rules will change etc). On the BRMS server, we need to have some background resource management where we control how many builds/tests are going on so that we dont' starve the BRMS as resources (so in that sense it becomes a bit like a CI server, where it detects changes and then builds int he background). Regarding your profiling so far - nothing seems that surprising - were there any areas that you think needed particular attention? Michael. On Nov 25, 2007 6:43 AM, Carlsen, Len <len.carlsen(a)ubc.ca&gt; wrote: -- Michael D Neale home: www.michaelneale.net blog: michaelneale.blogspot.com

Hi Len. Sounds like your requirements are pretty clear, which is great. In terms of scaling to that level we have the BRMS - which will need some (planned) improvements to make it possible to navigate such a huge amount of stuff without anyone going insane. On the otherwise we have the engine needing to cope with this onslaught of rules. It seems there there are several applications to consume the rules, the largest single one being based around courses (being 2000 courses with about 10 rules each, which sounds pretty sensible) - so that would probably be the most challenging thing in terms of memory use (its having enough memory available to build/load etc that mostly concerns me - that actual performance once you get past that should not really be an issue). Would like to talk about this more offline if you are willing (just cause there are lots of Q's that we have) - but it is interesting as cases like these don't come along all that often (and its hard to anticipate all requirements). Would also like to (if possible) take a look at what you are doing already (as you have the course rules working it sounds like) and build it from there. Please ping me if you want to set up a time to chat (on IRC, or we can just exchange mails, or voice etc). Michael. On Nov 9, 2007 9:06 AM, Carlsen, Len <len.carlsen(a)ubc.ca&gt; wrote: -- Michael D Neale home: www.michaelneale.net blog: michaelneale.blogspot.com