Implementing OTM Fleet Management in a Large-Scale Integrated Environment: Successes and Challenges Ted Gifford OTM SIG Conference Philadelphia PA August 15, 2012
Outline Schneider National Overview Overall Project Scope Go Live and Rollout SNI OTM Fleet Implementation Conclusion pg.2
Schneider is a transportation leader with a broad portfolio of services. Schneider National, Inc. is a premier provider of transportation, logistics, and intermodal services Operates more than 11,000 tractors, 15,000 drivers, and 55,000 trailers and containers with 22,000 associates. Intermodal TruckRail TruckRail Logistics Express Supply Chain Management International Logistics Transportation Management Truckload Long Haul Dedicated Regional Bulk pg.3
Truckload Operations 10,000 orders tendered daily to multiple modes Over-the-Road, Intermodal, Partner Carriers Long Haul (LH), Regional, & Dedicated Fleets LH Lead Time Distribution 2 hours to 14 days, Mean 60 hours Length Of Haul Distribution 200 to 4000 mi, Mean 700 mi LH Drivers/Tractors stay out 1-3 weeks Intermodal Drivers haul 2-4 loads/day Next load assigned 4-8 hours before Pickup pg.4
Outline Schneider National Overview Overall Project Scope Go Live and Rollout SNI OTM Fleet Implementation Conclusion pg.5
Project Scope Order Capture & Quote (Revenue Management) Asset Management & Maintenance Fleeting Planning, Rating & Costing Fleet Execution Driver Management Financials Oracle Fleet Management Components OTM Siebel CRM EBS HR EBS FIN SOA Middleware pg.6
Schneider Developed Components pg.7 Customer Rating Publication, Accessorial automation/billing, Rate Navigation Revenue Management Contract and Spot Pricing, Bid Evaluation Load Acceptance Capacity Availability, Service Reporting, Profitability Analysis Network Management Operational Forecasting / Demand & Supply Balance Short-Haul (load matching) Optimization Intermodal Dray & Regional Fleets Mobile Communications
Application Architecture Diagram pg.8
Outline Schneider National Overview Overall Project Scope Go Live and Rollout SNI OTM Fleet Implementation Conclusion pg.9
Go-Live and Roll-out Scheduled Driver Transitions by LOB 1400 1200 1000 800 600 VANS VANR RAIL DEDC 4404 2514 1371 3519 11,808 400 REMAINING 1,507 200 0 pg.10
Daily System Volumes Driver Assignments EOBR messages Order Releases Shipments Shipment Hub Tracking Events Vouchers Work Invoices Estimated 9,000 60,000 11,000 6,000 238,000 3,000,000 3,000 37,000 Actual 11,978 235,040 8,842 14,613 646,000 702,825 2,820 23,223 Agent Executions DB SQL calls 9.5 million 2.5 billion pg.11
Outline Schneider National Overview Overall Project Scope Go Live and Rollout SNI OTM Fleet Implementation Conclusion pg.12
SNI OTM Fleet Implementation Development of non-oracle components Code Extensions to OTM Automation Agents Integration & Component Interfaces Data Model extensions Functional Testing Performance Testing and System Scalability pg.13
Extensions to OTM Events Pre-processing & Handling Mobile communications External data inputs External Workflow Management of non-otm-native data statuses, references, remarks, special services. Closely-coupled companion Components Short Haul (Intermodal) Optimizer Trailer/Container Management Shipment & Driver HUB pg.14
OTM Implementation Automation Agents Number of Agents Shipment Tracking Event Order Release Driver Location Work Invoice Invoice Shipment Stop Shipment Special Service Equipment Other 446 196 169 21 17 7 6 6 5 4 4 11 Lines of Code 18,400 pg.15
OTM Implementation Automation Agents Lesson Learned Agent Framework is not a programming environment and should be used principally to automate OTM processes and NOT to extend product functionality. In particular, the framework lacks Robust constructs for conditional logic or looping Provision to define functions Debugging capability Control of synchronicity Ability to monitor error conditions Auto Retry failed actions Tracking individual agent actions pg.16
Integration & Component Interfaces To/From OTM Seibel (customer & order) Mobile Communications Dray Optimization Relay Optimization Trailer Management Data Hub Driver Compliance Mapping Dist/Geog Service Asset Maintenance ODS Analytics EBS HR & Fin Customs B2B Other 6 5 4 5 5 3 2 1 2 5 1 14 3 9 74 pg.17
Integration & Component Interfaces Implemented CDC (change data capture) via Data Hubs (shipment and driver) to provide timely OTM data to other subsystems. Initially planned to leverage ORACLE AIA PIPs now using only EBS-OTM driver PIP (driver data and work invoices) but plan to replace. Due to OTM and Oracle limitations, we developed structures to support Synchronous messaging protocol Unit of order for multi-threaded processes Locking and flexible queueing schemes Error handling and retry on failure pg.18
OTM Implementation Data Model extensions Additional data elements Order Release Shipment Special Services 5 3 References 30 50 Remarks 15 4 Status Indicators 43 90 Stop Special Services (per) 3 8 Stop Remarks (per) 3 1 Ship Unit References (per) 13 14 pg.19 Inefficient to query on these elements as they are stored in secondary tables as type-value pairs
Testing Number of Business Scenarios Tested Process Integration Final Claims 44 9 Generate Business 229 Acquire Customer Order 4162 1 AO-EDI 1110 411 Revenue Management 2231 46 Execute Order (OTM) 4941 47 STM 125 PP-Procure to Pay 522 38 Manage Receivables 543 97 General Ledger 328 82 Loss Prevention 237 22 Driver Pay 1120 196 HR-Hire to Retire 454 78 HR-OLM 301 7 QTC-Quote to Cash 77 Total 16347 1111 pg.20
1/17 1/21 1/27 2/2 2/8 2/14 2/18 2/24 3/2 3/8 3/14 3/18 3/24 3/30 4/5 4/11 4/15 4/21 4/27 5/3 5/9 5/13 5/19 5/25 5/31 6/6 6/10 6/16 6/22 6/28 7/4 7/8 7/14 7/20 7/26 8/1 8/5 8/11 8/17 8/23 8/29 9/2 9/8 Testing Retest Factor ~ 1.5, Automated Regression Testing Staffing 130 FTE Business, 30 TCE, 5 Automation Data 26K orders, realistic network configurations, etc. Defects order execution related 4,987 OTM SRs opened 847, fixed 447, ER 87 OTM C/H Open Defects by Severity 400 350 300 250 200 150 100 50 0 # Open High # Open Critical pg.21
OTM Physical Application Architecture Web Tier Web Tier Web Tier Web Tier UxPrd435 UxPrd436 Read Only App Tier App Tier Read Only App Tier App Tier App Tier App Tier UxPrd431 UxPrd432 UxPrd433 UxPrd433 OTM DB1 OTM DB2 OTM DB3 OTM DB4 OTM DB5 OTM DB6 pg.22
System Performance and Scalability 120 100 Changes July 7 Node Affinity Cache Adjustments Table Reorganization 80 60 40 20 Node 1 (RW) Node 2 (RW) Node 3 (RW) Node 4 (RW) Node 5 (RW) Node 6 (RW) 0 Buffer Gets (10^8) Disk Reads (10^6) SQL Execs(10^6) Cluster Waits (10^6) Total Wait Time (10^3 s) %DB time 140 120 After 100 80 60 40 20 Node 1 (RW) Node 2 (RW) Node 3 (RW) Node 4 (RW) Node 5 (RO) Node 6 (RO) 0 Buffer Gets (10^8) Disk Reads (10^6) SQL Execs(10^6) Cluster Waits (10^6) Total Wait Time (10^3 s) %DB time pg.23
2/13 2/20 2/27 3/5 3/12 3/19 3/26 4/2 4/9 4/16 4/23 4/30 5/7 5/14 5/21 6/4 6/11 6/18 6/25 7/2 7/9 7/16 7/23 Response Time (secs) System Performance and Scalability 110 Saved Query Weekly Average Response Time 100 90 80 70 60 50 40 30 20 10 0 Added DB CPUs "Re-used " Shipment Table Fields New Hardware - solid state disk DED IM TL Converted 2 DB nodes to read-only pg.24
Outline Schneider National Overview Overall Project Scope Go Live and Rollout SNI OTM Fleet Implementation Conclusion pg.25
Conclusion OTM Fleet can adequately support smaller fleets and limited integrations but needs additional development in key areas to support a large scale integrated environment... Improve robustness for rework, change, and imperfect data. Improve usability of Decision Control Center. Expand fleet data model & improve database efficiency (sql call ratio). Add fleet -related configuration screens and associated metadata. Improve technical and functional integration. Improve agent framework or provide alternative logic extension scripting capability. pg.26
Thank you Ted Gifford 920-592-5183 giffordt@schneider.com