Dynamic Conference Session Allocation Iron Architect Competition TechEd 2006, Boston Steve Land Principal Architect, Codesic Consulting.

Slides:



Advertisements
Similar presentations
Two Components IMAT consists of two components
Advertisements

Faculty of Information Technology Department of Computer Science Computer Organization Chapter 7 External Memory Mohammad Sharaf.
Shannon Bray Technical Architect Planet Technologies SESSION CODE: OSP310 Mike Watson Principal Consultant SeriousLabz.
Background Virtual memory – separation of user logical memory from physical memory. Only part of the program needs to be in memory for execution. Logical.
MODULE 3: OS & APP LAYERS. Agenda Preparing and importing a gold image Creating and understanding Install Machines Creating basic Application layers Understanding.
RAID Redundant Array of Independent Disks
P3 / 2004 Register Allocation. Kostis Sagonas 2 Spring 2004 Outline What is register allocation Webs Interference Graphs Graph coloring Spilling Live-Range.
Greedy Algorithms.
Register Allocation CS 671 March 27, CS 671 – Spring Register Allocation - Motivation Consider adding two numbers together: Advantages: Fewer.
CUPA-HR Presenter Preparation Presented by: Shiona Christensen CUPA-HR June 4-7, 2012.
Caching and Virtual Memory. Main Points Cache concept – Hardware vs. software caches When caches work and when they don’t – Spatial/temporal locality.
Iron Architect Competition Saravana Kumar Digital Deposit Ltd
 The IP address and port combination at which the NetScaler appliance receives client requests for the associated web application.  A public endpoint.
CS-3013 & CS-502, Summer 2006 Multimedia topics (continued)1 Multimedia Topics (continued) CS-3013 & CS-502 Operating Systems.
Scheduling with Uncertain Resources Reflective Agent with Distributed Adaptive Reasoning RADAR.
Tirgul 7. Find an efficient implementation of a dynamic collection of elements with unique keys Supported Operations: Insert, Search and Delete. The keys.
Hands-On Microsoft Windows Server 2003 Networking Chapter 7 Windows Internet Naming Service.
Systems Analysis I Data Flow Diagrams
Confidential and proprietary materials for authorized Verizon personnel and outside agencies only. Use, disclosure or distribution of this material is.
Conference Planning An ACEware Webinar. Course Setup ◦Planning.. Planning.. And more Planning ◦Fee Structure Name & Reg UDF’s  (finding space to store.
Machine Emulation & Developer Workstation Environment – Microsoft’s VirtualPC Dan Dyer Metrolist, Inc.
Cap 333 Network Administration. Grades  20 marks distributed on  Assignments / Project Activities Individual Pairs  1 or 2 tutorial quizzes.
Doc.: IEEE /0741r1 Report July 2015 Jon Rosdahl (CSR)Slide 1 1 st Vice Chair Report July 2015 Date: Authors:
Prof. Vishnuprasad Nagadevara Indian Institute of Management Bangalore
Virtual Memory.
Caching and Virtual Memory. Main Points Cache concept – Hardware vs. software caches When caches work and when they don’t – Spatial/temporal locality.
CSE 321b Computer Organization (2) تنظيم الحاسب (2) 3 rd year, Computer Engineering Winter 2015 Lecture #4 Dr. Hazem Ibrahim Shehata Dept. of Computer.
SYN407D: Image Management made easy with Provisioning Services 6.0
Page 19/17/2015 CSE 30341: Operating Systems Principles Optimal Algorithm  Replace page that will not be used for longest period of time  Used for measuring.
MS Tech-Ed 2006 Iron Architect Competition Greg Cogdell Milliken & Co.
TechEd Session Placement Efficient Space Allocation.
EXPOSE GOOGLE APP ENGINE AS TASKTRACKER NODES AND DATA NODES.
Terminodes and Sybil: Public-key management in MANET Dave MacCallum (Brendon Stanton) Apr. 9, 2004.
The EigenTrust Algorithm for Reputation Management in P2P Networks
Copyright © 2006 by Maribeth H. Price 2-1 Chapter 2 Working with ArcMap.
Department of Psychology Experiment Management System Experimenter Tutorial Stony Brook University Subject Pool Office
Cap 333 Network Administration. Grades  20 marks distributed on  Assignments / Project Activities Individual Pairs  1 or 2 tutorial quizzes.
Doc.: IEEE /0994r0 Report September 2015 Jon Rosdahl (CSR - Qualcomm)Slide 1 1 st Vice Chair Report Sept 2015 Bangkok Date: Authors:
Quark Matter 2009 Conference Guide. KCC, Middle (2 nd ) Floor Student Day + Parallel D: Lecture Hall Posters: Park Concourse + Room 200 A-C Cyber Cafe:
Copyright © 2006 Epsico, Inc. 1 Iron Architect Request for Proposals Dynamic Conference Manager 1.0 by Sandullah Epsicokhan Epsico, Inc.
Ted Pattison Author / Instructor Critical Path Training SESSION CODE: OSP315.
Hotel Registration A high contrast solution John Lomnicki ELCA Board of Pensions.
CS223: Software Engineering
1 An infrastructure for context-awareness based on first order logic 송지수 ISI LAB.
John R. Durant Senior Product Manager Microsoft Corporation SESSION CODE: OSP313.
COMP091 – Operating Systems 1 Memory Management. Memory Management Terms Physical address –Actual address as seen by memory unit Logical address –Address.
Competition Logistics 2016 Microsoft Indoor Localization Competition.
Building web applications with the Windows Azure Platform Ido Flatow | Senior Architect | Sela | This session.
#GPUGsummit Plan Ahead to Gain ROI During Summit!.
MS Office Project Portfolio Server Overview of Project Server Gateway.
Skype.
Protecting Your Data in 2017
Memory Management.
Microsoft Office Live Meeting 2007
Chapter 8: Main Memory.
Multiple Platters.
CE Courses at JSM and Other Meetings
Data Dissemination and Management (2) Lecture 10
Improvements in Outlook Calendar
Reserved Net Conference for Cisco WebEx Event Center
Communication Networks NETW 501
Edge computing (1) Content Distribution Networks
Multistep Processing of a User Program
Outline Module 1 and 2 dealt with processes, scheduling and synchronization Next two modules will deal with memory and storage Processes require data to.
SUSE Linux Enterprise Desktop Administration
Replica Placement Model: We consider objects (and don’t worry whether they contain just data or code, or both) Distinguish different processes: A process.
Lecture 3: Main Memory.
CSE 542: Operating Systems
Data Dissemination and Management (2) Lecture 10
Page Main Memory.
Presentation transcript:

Dynamic Conference Session Allocation Iron Architect Competition TechEd 2006, Boston Steve Land Principal Architect, Codesic Consulting

Problems Reduce conference walk time Allocate sessions based on popularity Distribute VPCs and slides in time for session Notify attendees of room allocation

Limits Solution should use Microsoft technologies

Assumptions Assume that we have awareness of individuals’ location at the room level of session rooms only  Not XY coordinates of each individual Assume that developer resources are available to build the solution

Context for dynamic allocation Not every attendee pre-registers Some attendees register for multiple events in a single time slot Speakers’ locations outweigh attendees in priority (setup, preparation) Travel between sessions matters less when the break is long, for example lunch breaks and expo hours

Dynamically allocating sessions The optimal solution would be for each person to stay in the same room However, it is not a burden to walk down the hall For any given room in the conference hall, one could create a list of second, third, and fourth preferred rooms relative to that room. For any given hallway in the conference hall, one could create a list of second, third, and fourth preferred hallways.

One (hu)man, One vote Actually one preference ranking Could be grouped by zones Zones could be grouped by larger zones (hallways) All these could be arranged in a hierarchy

One (hu)man, (more than) One vote 259AB 258ABC 257AB259AB 258ABC My room preferences, in order: 258ABC Either 259AB or 257AB 259AB 258ABC

Nested logical zones – simpler to capture than spatial data Zone3Zone4 Hallway Zone Zone1Zone2 Hallway Zone My zone preferences, in order: Zone3 Zone4 Zone2 Zone

Example starting data Floor1  Hallway1 Zone1  210ABC Zone2  206AB  205ABC  204AB  203  Hallway2 Zone3  259AB  258ABC  257AB This data would also need to capture room capacity. This example data is meant to show that basic room relationships can be expressed without spatial data. Avoids the need to create spatial map view for new venues.

Ranked voting Each individual “votes” for zones that are their top preferences for a given session, based on the room they are in. Basic approach based on Condorcet voting rules (ranked pairs) Allocate largest rooms to most popular sessions, next largest second, and so on Room preferences could be pre-generated for each room Speaker top-level zone and room should be considered first for room assignment

What about non-registered people? Algorithm based on approximate % total who attend sessions each hour Distribute attendee proportion across the most likely attendee count for totals

What about multiple sessions registered in one timeslot? Calculate probability based on other registration and frequency of users’ registration Randomly choose one session, weighted by probability, for each user

General outline approach for session allocation Tally all ranked votes Starting with most popular sessions, assign rooms with enough capacity based on votes  If instructor’s location is in top 10 of the ranked pair votes, use that location  If instructor’s second and third choices for location are in the top 10 of ranked pair votes, use that  Etc

Distributing VPCs and slides Central Web Service returns location of day’s bits  All VPCs for the entire session available in two or more central locations  Nightly process gets path and schedules download  Startup process verifies files are current  Morning and lunchtime verification re-verifies “Pull” process:  Each PC pulls the manifest from a Web Service  Automation deletes existing VPCs from local disk  VPC files transferred using locally-installed automation (Robocopy) Failsafe  Manifest offers at least 2 locations for download  Additional Virtual Server in case local copies fail

Presenter PC Daily bits WS Central File Store Request manifest XML file manifest Robocopy VPC files (all for the day) XML file manifest tells the PC automation the entire list of files to get and their paths. It also assigns a download start time that can be staggered Request manifest Verify files Nightly (after last presentation) Early morning, repeated at lunchtime [File discrepancy] Robocopy VPC files (deltas) Each presenter PC ends up with a local copy of all VPCs and slide decks for all sessions for that day Delete old VPC files

Logical deployment view – VPC and slides Get file location manifest Copy all session files locally

How to notify attendees Airports are a good analogy: dynamic gate assignments TechEd screens could show upcoming session locations in strategic locations  Current screens may not be seen; trusted source of info is still the booklet  Need enough room for everyone to view the screens Image source: Boston.com

Logical deployment view – Scheduling http LAN