Lecture 33: Scheduling and the Web © J. Christopher Beck 2008 1 Outline Using the Web to Solve Scheduling Problems Web-based information infrastructure Scheduling services Scheduling Web-based Processes GRID Scheduling Workflows © J. Christopher Beck 2008 2 Readings P Ch 14.6 © J. Christopher Beck 2008 3 The Information Problem Last lecture, we discussed the information problem for scheduling You need the right, dynamically updated information Does the web help with the information problem? © J. Christopher Beck 2008 4 The Current Picture Suppliers The rest of the information system Factory floor Customers © J. Christopher Beck 2008 5 The New Picture? Competitors Sales Customers Forecasting Suppliers Marketing © J. Christopher Beck 2008 Factory floor Shipping Inside Outside 6 The New Picture? The scheduling system can dynamically pull up-to-the-second information from all over (and beyond) the enterprise Can always schedule with the best available information and can reschedule independently © J. Christopher Beck 2008 7 Challenges for The New Picture The existence of the information Is the information in a computer system or in someone’s head? The form of the information Is it just textual or randomly represented or in some machine “understandable” format Ontologies and the semantic web Standards © J. Christopher Beck 2008 8 Challenges for The New Picture The quality of the information Business processes may not support updating the information How would you set up the business process to gather accurate data on the processing time of a given operation? Information Capture Problem © J. Christopher Beck 2008 9 Challenges for The New Picture Information integration You call it “activity”, I call it “operation”, they call it “task” – is it all the same thing? How do we automatically combine information from databases that were independently created for different purposes Shop floor system, customer-relationship management, marketing, … © J. Christopher Beck 2008 10 Interoperability Process Modeler Process Planner (ProCAP / KBSI) (MetCAPP/Agiltech) Scheduler Simulator (Quest / Dessault) © J. Christopher Beck 2008 (ILOG Scheduler) 11 Challenges for The New Picture Automated Reasoning If we mount a marketing push, we will increase orders How do we automatically reason about the implications of individual and combined information? © J. Christopher Beck 2008 12 Challenges for The New Picture The Information Firehose Be careful what you wish for Imagine all the information is available: how do you find, filter, recognize the information that is important for your task? Like trying to drink from a firehose © J. Christopher Beck 2008 13 Challenges for The New Picture The role of and interface for the user There will always be something that the user knows that isn’t represented Is this true? Does the human still bring value to the scheduling process? What style of interaction do we provide? User is a source of information? User can change the schedule? Who has decision making authority? © J. Christopher Beck 2008 14 The New Picture? Competitors Sales Customers Forecasting Suppliers Marketing © J. Christopher Beck 2008 Factory floor Shipping Inside Outside 15 The Newer Picture? Competitors Sales Forecasting Customers © J. Christopher Beck 2008 Suppliers Marketing Factory floor Shipping Inside Outside 16 The Newer Picture? Service Oriented Architecture The “nodes” on the web aren’t just databases – they provide services, reasoning, and information © J. Christopher Beck 2008 17 The Newer Picture? Competitors Sales Forecasting Customers © J. Christopher Beck 2008 Suppliers Marketing Factory floor Shipping Inside Outside 18 The Newer Picture? Competitors Sales Forecasting Customers © J. Christopher Beck 2008 Suppliers Marketing Factory floor Shipping Inside Outside 19 The Newer Picture? Competitors Sales Forecasting Customers © J. Christopher Beck 2008 Suppliers Marketing Factory floor Shipping Inside Outside 20 The Newer Picture? Competitors Sales Forecasting Customers © J. Christopher Beck 2008 Suppliers Marketing Factory floor Shipping Inside Outside 21 The Newer Picture All the challenges (and more) of the “new picture” apply here But there is an example of something like this in the high performance scientific computing world: The GRID © J. Christopher Beck 2008 22 The Computational GRID Observation Data Visualization Experimental Data Data Storage © J. Christopher Beck 2008 Data Processing 23 The Computational GRID Computing services are available to the scientific community High-energy physics, astrophysics, computational biology, etc. Services: Number crunching, visualization, data storage (terabytes of data!) © J. Christopher Beck 2008 24 The Computational GRID Reality Much of it is still point-to-point The user needs to organize the interaction of machines to get the desired functionality But people are working on all the challenges © J. Christopher Beck 2008 25 Scheduling the Web With a service oriented architecture, you can offer a scheduling service e.g. rental car reservation scheduling, transportation scheduling, … But, you also have the problem of “scheduling” (i.e., coordinating) a set of services to form some business process © J. Christopher Beck 2008 26 Scheduling the GRID You want to pull data from specific databases, run specific transformations and combinations, and visualize it Ideally, you’d like to specify this at a high level and have automated planning and scheduling tools take care of it © J. Christopher Beck 2008 27 Scheduling the (New) Web? UofT wants a special issue document for its 200th anniversary (in 2027) Automatically create and schedule the process find articles, find photos, select them, do the design layout, printing, mailing (find addresses), … What parts will humans have to do? © J. Christopher Beck 2008 28