IBM Web Sphere Transaction Cluster Facility Product Evaluation

IBM Web. Sphere Transaction Cluster Facility Product Evaluation Summary www. tpfsoftware. com

Agenda Ø Product Definition / Architecture (IBM) Ø Background - TPF Software with WTCF Ø Discussion – Application development process Ø Hands-on - Sample Applications Ø Sharing - Our “First-look” Experience Ø Final Notes / Comments Ø In Conclusion / Future Plans 2/21/2021 TPF Software Inc. 2

Product Definition / Architecture 2/21/2021 TPF Software Inc. 3

IBM Web. Sphere Transaction Cluster Facility Bringing extreme transaction processing against shared data to distributed systems § Born out of IBM Research and based on high volume transaction processing concepts from IBM System z § IBM Web. Sphere Transaction Cluster Facility (WTCF) § Provides an optimized offering for mission critical shared data OLTP § Built on the latest DB 2 shared data technology § Supporting workloads that require: § Consistent speed and high transaction throughput with 24 x 7 availability § Scalability to grow, incrementally and smoothly, with your business § Shared, always consistent, view of mission critical information required to process transactions Workload optimized for performance, efficiency and scalability § WTCF can be well suited for a range of workloads: reservations, payments, risk, fraud, cargo, logistics, traceability (RFID)… § WTCF is a bundled offering that includes: § IBM Web. Sphere Application Server § IBM DB 2 Enterprise Server Edition § IBM DB 2 pure. Scale feature for Enterprise Server Edition 4 § WTCF Tivoli Monitoring Agent Web. Sphere Transaction Cluster Facility • Announced: 9/20/2011 • Available: 10/7/2011

Software Web. Sphere Transaction Cluster Facility is… Application Infrastructure Data Management & § IBM Web. Sphere Transaction Cluster Facility (WTCF) has the features and capabilities for creating highly scalable centralized/local transaction processing applications § WTCF concept initiated/prototyped by IBM Research, productized through close collaboration between multiple IBM divisions § WTCF builds on IBM’s other leading software components: ü WTCF applications can be written in Java, and deployed to WAS, or in C++ ü Built on AIX and Power Systems ü Built on DB 2, which provides the data store and manages/maintains the underlying storage ü Exploits the latest DB 2 cluster capabilities (currently called DB 2 pure. Scale) to provide high availability and increased scaling 5

Software WTCF – How it fits into a general application architecture WTCF Tivoli Monitoring DB 2 pure. Scale WAS WAS WTCF Tivoli Monitoring DB 2 pure. Scale WAS WTCF “applications” are WAS applications that are co-located on clustered servers where the data is managed Power WAS Power 6 WTCF is an offering targeted logically between the data/resource and the business logic tiers for a select workload Unique data management API to enable the creation of network model databases

Software Architected for High Volume, Centralized Transaction Processing Object. Oriented Multi. Tenancy Tenant 1 Data Integrity Developer Productivity 7 Tenant 2 Tenant 3 WTCF Database Record “A” Record “B” Heterogeneous Data Record “C” Network Model

Software WTCF Components Configuration Assistant Generated DB Classes for Java WTCF Java Library Web. Sphere Application Server Generated DB Classes for C++ Tivoli Monitoring Java Applications C++ Applications Monitoring Agent WTCF C++ Library Persistence Server DB 2 8 Shared Main Storage

Background - TPF Software and WTCF 2/21/2021 TPF Software Inc. 9

Genesis �IBM Labs offered pre-release look at WTCF �TPF Software accepted opportunity: �Learn about new technology �Benefits for existing products, services and customers �Opportunities for new products or services 2/21/2021 TPF Software Inc. 10

Training, Set Up & Support �Training �IBM Labs provided 2 -day online training �Development System �Partner World Virtual Loaner Program (VLP) system � IBM POWER 7™ server with IBM AIX® �Installation �IBM installed and updated WTCF on VLP system �Help & Support �Weekly checkpoint calls �Special help with setting up paths & profiles, database design, node connections/restarts, and examples & samples �Normal 14 -day VLP renewal period extended to 90 days 2/21/2021 TPF Software Inc. 11

WTCF – Product Evaluation & Study �TPF Software started “WTCF Product Evaluation & Study” �Plan: �Induct a team to work on sample assignments. �Formulate evaluation process – with hands-on experiments. �Define evaluation perspectives & criteria. �Design sample exercises to test criteria. �Record results. �Re-iterate steps based on observations / changes / maturity. �NOTE – Team has had only 3+ months of hands-on experience with WTCF. Not all steps of project plan have been carried out. 2/21/2021 TPF Software Inc. ! 12

WTCF Application Development Process 2/21/2021 TPF Software Inc. 13

WTCF – Database design � Application databases are represented as a hierarchy of database objects – { Fileroom -> Cabinet -> Folder -> Record -> Fields } � Database Configuration Assistant in WTCF toolkit is used to design most of the database entities: � Tenants, Fileroom, cabinets, records, fields, user & managed references. � Folder Indexers, Folder Locators, User defined folder distribution algorithms. � Named row collections, DB 2 tables etc. � Auto generate programming artifacts. 2/21/2021 TPF Software Inc. 14

WTCF – Application Development Process 1 DB design using configuration assistant Compile the application – include static libraries in the path 8 2 3 Save the database configuration as [. wtcf] file. Auto generate application classes from the saved DB configuration file. 7 6 Include the autogenerated headers into application sources. 9 Resolve external references to the fileroom by linking to the shared library. 2/21/2021 Run makefile to generate the fileroom library shared object. 4 Transfer the configuration file and programming artifacts to AIX machine 5 Setup the WTCF library paths. 10 Application is ready for execution. TPF Software Inc. 15

WTCF – Using the new Database Allocate DB 2 resources 2/21/2021 • Create DB 2 TABLES & SCHEMAS • Create TABLESPACE & BUFFERPOOL for new WTCF database. Configure WTCF database • Create database configuration file using configuration assistant. Initialize WTCF database • Use wtcf_dbcfg -- init to initialize the database & node cluster Node Start • Start the Node using wtcf_nodestart command. Initialize root cabinets • Use wtcf_initcab to initialize the root cabinets. TPF Software Inc. 16

Hands-on – Sample Applications 2/21/2021 TPF Software Inc. 17

Hands-on experience �Started self learning with “FLIGHT” application – provided by IBM with WTCF installation. �Planned and designed several exercises to study / evaluate several database criteria: � Usability – Flexibility, robustness, suitability for complex database design, multi-level & multi-path databases, ease of implementation, etc. � Offered features – Auto code generation, single window design, complex query handling, hierarchy, references, etc. � Advanced features – Multi-tenancy, node uniqueness, OLAP vs OLTP co-existence, Transaction scoping, Data integrity, concurrency, persistence extent, object vs page level locks, performance, disaster recovery, scalability, cluster management, Analytics, web/transaction management etc. Work in Progress!!! 2/21/2021 TPF Software Inc. ! 18

Sample 1 : Address book Application �Objectives – Study a very simple WTCF Application �Problem statement – Design an application called “address finder” – with following features & search facilities: � An Address record will contain following fields: � Names (1 or more) – First Name + last Name – STRING padded by space � Street Address – STRING padded by space � Apt/bldg/unit number (can be null) – INT � city name – STRING padded by space � State – STRING padded by space � Zipcode – LONG � Possible user queries: � Retrieve list of matching addresses based on first 3 chars of last name. � Retrieve list of matching addresses based on zip code � Retrieve list of matching addresses based on [number of people] – less than or equal to or greater than a number + first 3 chars of last name (Manual matching) � Retrieve list of addresses based on Name + city – (Manual matching) 2/21/2021 TPF Software Inc. 19

Data Model for Address book 2/21/2021 TPF Software Inc. 20

Sample 2 : Passenger Check-in Database �Objectives – To study a complicated WTCF application �Sample Application – Passenger Check-in data store �Specifications: � Each passenger folder will have Name, Seat, Ticket, PNR, SSRs, Documents (PASSPORT ETC), Security, Baggage records � One passenger data per folder (detail file) � Passenger can have multiple bags. � Multiple passengers can have same names. � Passenger can have multiple SSRs. � Passenger can travel without a PNR. � Passenger can be re-accommodated from one flight to another. � And other rules & relationships wherever applicable. 2/21/2021 TPF Software Inc. 21

Passenger Check-in Database (Existing) 2/21/2021 TPF Software Inc. 22

Passenger Check-in Database (WTCF) 2/21/2021 TPF Software Inc. 23

Passenger Check-in Database – Analysis Cabinet Type References Folder Locators Folder Indexers Records Name. Cab Root Passenger Nameby. Last. Name N/A Name. Rec Pnr. Cab Root Passenger PNRby. Ref N/A Pnr. Rec Docs. Cab Root Passenger Docsby. Num N/A Docs. Rec Tkt. Cab Root Passenger Ticketby. Num N/A Tkts. Rec Bags. Cab Root Passenger Bagby. Num N/A Bags. Rec City. Cab Root Days. Cab Cityby. Code N/A City. Rec Days. Cab Mid Flt. Nbr. Cab Daysby. Citydate Daysby. City. Date. Inde xer Days. Rec Flt. Nbr. Cab Mid Passenger Fltby. Citydate. Fltnbr Fltby. City. Date. Fltnb r. Indexer Flt. Nbr. Rec Passenger Leaf N/A Paxby. Name Paxby. PNRREF Paxby. Docs Paxby. Tkt Paxby. Bag Paxby. Citydate. Flt Paxby. First. Name. Last Name. Indexer Paxby. PNRREFIndexer Many 2/21/2021 TPF Software Inc. 24

Summary - Hands-on exercises E. No Title Status Remarks 1 Address book Development Learning database design 2 Pax-CKIDB Development Multi-level indexing 3 Availability – POS application Design Multi-level indexing & accesses / updates 4 Message Box Design Concurrency, transaction volumes, data integrity. 5 Message Manager Design Fast persistence, Lock & release mechanisms, Data Integrity & Volumetric tests, data import/export mechanisms, complex mapping analysis etc. 6 Booking Engine – Core Concept OLAP & OLTP – coexistence, STAR schemas designed for warehouse queries, query complexities. 7 ---- 2/21/2021 TPF Software Inc. 25

Sharing - “First-look” Experience 2/21/2021 TPF Software Inc. 26

Some observations Database Design • Hierarchy + Relational database mixture makes it easy to handle complex designs. • ADBI abstracts the database operations. • Multi-level cabinets accommodates multiple access / update paths. • Managed/ user references make parent-child relationships easy. • Multi-tenancy/multi-hosting is an inherent feature – makes it very simple to implement. 2/21/2021 Application Development Configuration Assistant • Auto code generation makes it easy to use the Database objects. • Smart pointers – Address and arrest memory leaks. • More than 40 p. c. of code is auto-generated. • Easy to re-engineer applications • ADBI provides very simple & straight-forward interaction facility. • Captures entire database design into a single metadata file. • Facilitates auto code generation • Helps in validating the database metadata • GUI based – Makes it very easy to use • Allows single window database design. TPF Software Inc. 27

Final Notes / Comments �WTCF + WAS + AIX – Can offer a powerful application processing platform for applications requiring: �Common database for OLAP & OLTP �Distributed infrastructure �Horizontal scalability without single point of failures �Single point of control �Multi-tenancy & Clustered implementations. �Default support for Object oriented programming paradigms without any wrappers or adapters. �Seamless failover switching for Node failures & Cluster failures. �Simple & visual aided database design. 2/21/2021 TPF Software Inc. 28

In Conclusion �Learn about new technology �Benefits for existing products, services and customers �Opportunities for new products or services 2/21/2021 TPF Software Inc. 29

TPF Software – Future plans �Additional products for WTCF – data loaders, single point of control tool for nodes & node clusters, import / export data features, etc. �Tools for WTCF – development studio, auto code generation engines, debuggers, etc. �Continue evaluation and write white-papers. �Work with IBM on any go-to market strategies or future enhancements. �Work with customers to get hands-on experience. 2/21/2021 TPF Software Inc. 30