Bryan Tuft Sr. Sales Consultant Global Embedded Business Unit [email protected] Agenda

• Oracle Approach • Embedded Databases • TimesTen In-Memory Database • Snapshots • Q&A Real-Time Infrastructure Challenges

Very high data throughput rates – Accelerating… • Joint Operations across Services and Allies • Global Information Grid • RFID & other Sensors Everywhere • Wireless Communications

Flexible Infrastructure Requirement • Standards • Strong Desire for COTS • Service Oriented Architecture Oracle’s Approach…

• Secure • High Performance • Standards Based and Interoperable • Based completely on “COTS” products • Ease of integration (embeddable) Acquisitions Make the Oracle “Stack” More Compelling Open Source

Consulting Services Oracle Consulting IBM GlobalServices SAP Consulting 3rd Party SIs 3rd Party SIs

Microsoft SugarCRM, Applications Fusion 3rd Party ISVs mySAP Business others Applications Solutions

Development Tools Oracle Tools / J2EE Rational / WebSphere NetWeaver / ABAP Visual Studio .NET PHP, Eclipse

Middleware JBoss, Spring, Suite Fusion WebSphere NetWeaver .NET Middleware Geronimo

Database Database 10 g, Oracle, MySQL, TimesTen IMDB, DB2 DB2, MySQL, SQL Server PostgreSQL Berkeley DB SQL Server

Operating //Windows Linux/Unix/Windows Linux/Unix/Windows Windows Linux System Oracle’s Embedded Data Management Oracle Lite Berkeley DB TimesTen

Oracle Lite Laptops / Database Appliances / Devices

Oracle Lite Application Mobile Tier Server

Oracle Lite Data TimesTen Cache Synchronization Connect to Oracle Update Propagation Which Embeddable Database?

• Speed! Microsecond latency and 100,000+ TPS • Cache synchronization with Oracle DB in real-time • SQL, ODBC, JDBC • Embedded or client-server operation • In-Memory database for ultimate performance

• Low latency and high throughput • Embedded only operation • Non-relational queries, very small footprint (< 200kb) • XML/XQuery capable

• Synchronization to (periodic) • Very small embedded footprint (<200kb) for handheld devices • SQL & Stored Procedures

• Performance, Reliability, Security and Portability • Handles any possible SQL or data type such as Spatial 10g • Footprint is not a constraint Oracle TimesTen In-Memory Database Proven in Real-Time Deployments Over 1,500 companies worldwide use Oracle TimesTen

In Networks In Telecom On Wall Street In the Enterprise

Real-Time Billing Value-Added Services Order Matching Call Centers

Voice over IP Revenue Assurance Risk Management Business Intelligence

Mobile Networks Network & QOS Mgmt Real-Time Analytics SOA’s In-Memory Database Technology

• Entire database is in-memory • Memory optimized indexes and algorithms: 10x CPU efficiency • Choice of embedded or client/server options • Standard APIs (ODBC, JDBC, SQL) • Flexible durability options for data persistence to disk • Embedded class ease-of-use • Excellent data replication for high In-Memory Database  Relational, standard API’s availability  Very fast response  Very high transaction rates  Enterprise scalability & reliability Finding a Record in TimesTen

Application Determine Memory Address of Desired Record

Query Optimizer Memory Address SQL /Executor

Memory-Resident Database Copies data to application buffers

The full database Data Store is preloaded from disk to memory TimesTen Throughput IBM eServer pSeries 670, 12 CPU 1.1 Ghz POWER4 AIX 5.2, 64bit Lightning Fast Response

Oracle TimesTen Release 6.0, 4-CPU, 3 GHz x86 Xeon, 32-bit RHLinux 30 28 20 millionths 25 of a millionths second of a second 10

microseconds 9 millionths 0 update mixed select 70% select; 20% update; 10% insert TimesTen-to-TimesTen Replication

• Replicate entire databases or individual tables Master - Subscriber • Dynamic configuration with SQL • Automatic recovery and catch-up of down systems Master - Master • Asynchronous or synchronous (dynamic configuration with SQL) • Does not include a cluster manager (no automatic N-Way failover) ACID Properties • Atomicity: full support for transactions (COMMIT/ROLLBACK) • Consistency: database is always in a consistent state, and recovers from disk in a consistent state • Isolation and locking: • Row-level locking • Thread safe • READ COMMITTED isolation • Writes do not block reads • Reads do not block writes • SERIALIZABLE isolation also available • Lock timeouts and deadlocks- very rare • Durability: • Transactions logged to disk and recovered Three Logging Types Dynamically configurable with SQL statements

LOG TimesTen BUFFER In-Memory ….. Tx …... Data Store ……Tx …… Logging Hard Disk …...Tx…… Hard Disk

Datastore.ds0 LOG Buffered TimesTen BUFFER Logging Datastore.ds1 ….. Tx …... Data Store ……Tx …… …...Tx…… Datastore.log1 Datastore.log2 LOG TimesTen BUFFER Datastore.log3 COMMIT Data Store COMMIT Durable COMMIT Commit Cache Connect to Oracle Application-Tier Data Access

User Apps Key Takeaways… JDBC – ODBC – SQL92 • Pre-load or load-on-demand the most active data from Oracle • Synchronizes updates in both directions • Transparent SQL pass-through for non-cached data requests • IMDB keeps working even if the connection to Oracle is down Snapshots & Case Studies TimesTen becoming widespread in DoD/Intelligence community

• TimesTen being used or evaluated by many Defense/Intelligence customers.

• How is it used? • Command and Control • Large memory decision support systems • Data Fusion • Data Distribution and Messaging • SOA • Application Tier Caching • Embedded DB in Real-Time Applications RTI - Distributed Data Service (DDS)

•RTI(Real-Time Innovations) is the leading real-time messaging vendor

•Customers include U.S. Navy, U.S. Air Force, U.S. Army, Northrop Grumman, Raytheon, Lockheed Martin, BAE Systems, etc.

•DDM scales to hundreds of nodes App App SQL/ODBC SQL/ODBC App •Publishes data from one to many S Disk •Used on ships and battlefield S operations control systems S DDS DDS DDS •DDM Embeds Oracle’s IMDB DDS so Real-time database nodes DDS DDS DDS can be distributed everywhere S App App SQL/ODBC SQL/ODBC S App App Disk Example: Joint Forces

• Proof of Concept with live data • Integrate Army, Air Force and Navy Systems • Pull data from 3 different interfaces • SOAP/XML • Direct Flat File Interface • Spreadsheet • Requirement: Application-Mid-Tier must function even when disconnected from back- end Database Data Fusion for Combatant Commanders

Application Server 10g • J2EE Container FALCON WEBSKED ARTIMS • Portal • Forms and Reports • Business Intelligence

BPEL Process Manager

In Memory Database

Enterprise Identity Management • Authentication & Authorization • Provisioning • Federation

Web Service Manager (Security)

MapViewer Joint Forces - Analyze

FALCON WEBSKED ARTIMS Access XML File Spreadsheet

Heterogeneous Service

J2EE • JSR168 Portlets • Data Flush BPEL • Business Intelligence

Linux TimesTen Data Cache Discoverer

Web Service Manager Portal END