400 likes | 549 Views
Scaling Applications Cost -Effectively to Meet Mobile and Cloud Demands. Venkata Ravipati Senior Director, Product management.
E N D
Scaling Applications Cost-Effectively to Meet Mobile and Cloud Demands Venkata Ravipati Senior Director, Product management
The following is intended to outline our general product direction. It is intended for information purposes only, and may not be incorporated into any contract. It is not a commitment to deliver any material, code, or functionality, and should not be relied upon in making purchasing decisions.The development, release, and timing of any features or functionality described for Oracle’s products remain at the sole discretion of Oracle.
Program Agenda • Introduction to Coherence • Customer Case Studies • Summary
Problem Domain of Coherence High load, stateful, middle tier services using shared services • Service Clients • Browsers, supplier and partner clients, application clients, mobile apps, and other service consumers. • Grow and scale naturally • Facing rapidly increasing load • Commonly stateful • Highly dependent on shared services Service Implementation Custom applications, BPM processes, service bus endpoints, reporting and batch services … • Complex and expensive to scale • High latency SPOB and SPOF Shared Services RDBMS, cloud services, supplier and partner services, mainframe applications, and so on.
Coherence In Memory Data GridScaling, Performance and Reliability of Application Infrastructure Applications and Middleware • Application Scalability • Reducing backend load with in memory caching – database, applications, mainframe • Application Performance • In memory data access – information close in memory, not in slow backend systems • Application Reliability • Reliable in memory data grid providing high-availability to applications Applications and Middleware Scale, Increase Performance &Improve Reliability with Peer to Peer, High AvailabilityIn Memory Data Grid Coherence
Coherence Clustered Caching ExplainedPartitioned, Fault Tolerant, Self-Healing Cache • Cluster of nodes holding % of primary data locally • Back-up of primary data is distributed across all other nodes • Logical view of all data from any node ? • All nodes verify health of each other • In the event a node is unhealthy, other nodes diagnose state • Unhealthy node isolated from cluster • Remaining nodes redistribute primary and back-up responsibilities to healthy nodes
Example Uses Coherence In-Memory Data Grid • Caching • Client applications request data objects and state from the Data Grid rather than backend data sources • Analytics • Client applications ask the Data Grid questions from simple queries to scenario modeling • Compute • Applications execute compute – business logic and transactions - across the Data Grid • Events • Applications take action based on events from the Data Grid
Oracle Coherence In-Memory Data GridSecret Sauce for Scaling, Performance and Reliability at Internet Scale Oracle acquires Tangosol Deep adoption in Fusion Middleware, Oracle Horizontal and Vertical and Business Applications and Exalogic 2000 2001 ….. 2007 2008 2009 2010 2011 2012 2013 Alignment with FMW Release Cycle Tangosol founded to pioneer distributed caching Coherence 1.0 Cool Vendor Award Coherence 3.5 Coherence 3.6 Coherence 3.7 Coherence 3.7.1.1 Coherence 12.1.2 Increasing market share and feature capabilities for In Memory Data Grid. Market widened All Categories – Financial Services, E-Commerce, Travel, Insurance, Retail, Manufacturing, Telco …
Delivering Next Generation ApplicationsSocial, Mobile, “Internet of Things” Demand New Architectures More Data from More Sources Users and Applications Demand Real-Time Updates Coherence: In-Memory Scalable Caching and Processing
Scaling Enterprise Resources Reduce requests for information and costs • Optimizing Data Access • Batching write operations and coalescing updates (write-behind) • Only refreshing stale data in the cache from a source database, using the up-coming “HotCache” feature (based on GoldenGate) • Maintaining Performance • Serving data requests from memory rather than from the Enterprise System
Scaling Enterprise ResourcesDelta Airlines Key Problem Customer Solution Key Benefits • Grow online presence while maintaining a great user experience • Manage transactions at peak times without impacting other systems • To contain costs • Cache read-mostly data in Coherence • Bringing data closer to application logic • Caching and replicating transitory data, like HTTP Session State • Reduced load on databases and other backend systems • Improving scalability of application tier without having to scale enterprise resources • Making it easier to predict growth costs
Accelerating Service Based Architectures Reduce requests for information to reduce costs • Improve service performance and scalability by caching responses • Control costs by reducing the requirement for additional system investment • Increase service availability
Accelerating Service Based ArchitecturesLarge European Telco Key Problems • Service enable legacy systems • Minimize additional investment in legacy systems • Provide a flexible and scalable solution Customer Solution • Cached legacy systems service responses • Throttling requests to minimize load • Replicated responses between data centers Key Benefits • Provided fast and scalable services • Scaled legacy systems without additional investment • Provided reliable service and operational flexibility
Improve Operational Efficiency Reduce maintenance windows and overheads • Reduce downtime with on-line Web Application patching and hardware upgrades • Enable maintenance in office-hours with zero disruption for users/customers • Provide inter-site replication for business continuity
Improve Operational EfficiencyON24, a Leading Global “Virtual Events” Provider Key Problems Customer Solution Key Benefits • Reduce maintenance costs • Improve application performance • Ensure high quality un-interrupted service • Used Coherence to seamlessly handle planned and unplanned outages • Shared data between applications server instances so each instance could be tuned for specific application • Utilized a fast and reliable in-memory cache • Used Coherence*Web to share HTTP sessions • Enabling online maintenance without disrupting users • Application specific tuning for optimal performance • Dynamic resource allocation to manage peaks in demand
Reduce Mainframe Costs Reduce Mainframe Interaction to Save MIP’s • Cache frequently accessed semi-static data, refreshing periodically • Cache dynamic data, synchronizing with a source database using the new “HotCache” feature, based on Oracle GoldenGate. • Improve cost/value ratio for data
Reduce Mainframe CostsAvis Car Rental Key Problems Customer Solution Key Benefits • Business logic and data on Mainframe (MF) • Cost growing exponentially when MF accessed from customer applications • Access frequently used data from a Coherence cache to reduce Mainframe MIPS • Simple use case with quick ROI • Avis reduced MIPS by 32%
Reduce Mainframe CostsMajor European Retail Bank • Easily Measurable ROI • “Badly” behaving applications accessed MF 10-30 with each user session • Customer related data accessed ~10M times per month • Customer Service MIPS reduced by ~40% • Solution had ROI of 2 months! – which included hardware • “Coherence is rock solid, over two years in production and not a single problem related to Coherence” – customer quote
Deliver Real-Time Information Push Information in Real-Time to Client/Customer Applications • Allow users/customers to see changes to data as they occur • To enable better decision making • Enable diverse web applications • Provide a better customer experience • Optimize system automation
Delivering Real-Time InformationSporting Index, an online sports spread-betting provider • Real-Time Event Hub • Coherence used as event hub, capturing pricing updates and delivering these in real-time to analysts using Excel • The event processing platform improved the utilization of the existing infrastructure, scaled their computing power 400% and grew revenue 15% in 6 months
Credit Suisse Delivering Real-Time InformationCredit Suisse and real-time credit checks Key Problems Customer Solution Key Benefits • Needed very low latency response times for pre-transaction credit checks against counter-parties globally • Process intraday credit checks for a large number of transactions daily and scale by up x10 without risk of increase in latency • Minimize solution development, support time, and cost • Chose Coherence for its simplicity, which enabled a team of 4 to deliver the system quickly and support it globally • Fully prototyped the system in 3 months • Used diagnostics in Oracle JRockit Mission Control to pinpoint performance bottlenecks • Achieved sub 5ms pre-transaction credit checks against counterparties for 99.9% of transactions • Globally replicated changes • Provided instant system failover
Delivering Real-Time InformationJP Morgan provides real-time risk and P&L for better decisions • Real-Time Risk and P&L • Coherence chosen as a resilient, highly available platform to provide a real-time view of critical Risk and P&L information • Cluster has scaled to 200+ JVM’s • Easily integrated with .NET client for the UI • Proved very flexible and able to meet JPMC’s specific requirements
Delivering Real-Time InformationEuropean Rail Operator • Real-Time Train Monitoring • Coherence used to handle 100’s of messages p/s, provide real-time feedback to operator screens in <=1s after after trains pass sensor • Critically it was still operational even after tests shutdown 50% of the grid • Solution resulted in more accurate operational data, better forecasting, reduced man-power and reduced the time to resolve traffic disruptions
Delivering Real-Time InformationMajor APAC Manufacturer • Real-Time Manufacturing • Coherence and OEP used to monitor manufacturing process, receiving up to 800 events p/s • Each event triggers calculation based on a sophisticated mathematical model • End-to-end transaction latency was under 20ms 99.999% of time
Fast Data For Internet Scale Process Big Data in Real-Time using Coherence with OEP and NoSQL • Filter, analyze and process real-time data using • Oracle Event Processor to filter and aggregate data streams • Coherence cache, lookup, store and share data horizontally • Oracle NoSQL to capture and archive
Fast Data For Internet ScaleTurkcell, the leading mobile provider in Turkey Key Problems Customer Solution Key Benefits • Needed a highly available and scalable complex event processing and data grid solution that could support a huge volume of network data and capture real-time streaming information from mobile networks. • The solution is deployed at Turkcell, Europe’s third and the world’s 16th largest mobile operator, which handles more than 33 million subscribers and 2 billion events daily • Chose Coherence to provide scalability and reliability for high-performance event-driven applications. • Able to identify and predict trends based on real-time subscriber events from mobile devices with SMS text message temporal relevance, to provide customers with a wide range of targeted services and promotions – when and where they are most relevant.
Fast Data For Internet ScaleOracle Elastic Charging Engine (ECE) • Fast Data Processing • Oracle Coherence and NoSQL used for processing and storing real-time service charging information • ECE can process over 3 billion events per hour with an average latency of 5ms, equating to more than enough processing power to rate and charge for all of the world’s phone calls on a single rack Exalogic machine
Fast Data For Internet ScaleA leading sports and leisure retailer/manufacturer Key Problems Customer Solution Key Benefits • Needed a highly available and very scalable platform to capture and process information from “sports devices” • Coherence used hold and process information for a Sports Platform • Data comes from a range of devices • 1.2B reads and 300M updates each day on 2TB of data • Coherence chosen for low-latency, HA, in-grid event processing and linear scalability • Linear Scalability • Low Latency Access to Recent Data • Event Driven computing on the Grid • High Availability
Fast Data For Internet Scale NTT Docomo and real-time mobile traffic processing Devices Mobile Web • Largest mobile operator in Japan • Oracle Event Processing and Oracle Coherence • Enriched and correlated traffic for Big Data • 13 million Smartphone users • Over 700,000 events/sec • 50% total cost reduction • Managing Events in Real-time Traffic Big Data Real-time traffic processingOracle Coherence & Event Processing
Integrating Web Applications Share Web Session State to Integrate, Partition and Optimize • Integrate Web Application By Sharing (HTTP Session) State • Unify separately developed applications for a seamless customer experience • Modularize and Migrate Web Applications • Partition large monolithic applications • Share session state between old and new applications • Combine Web Technologies, like JEE Applications and Portals
Integration Web ApplicationsMajor Global Retailer • Application Consolidation • A global retailer needed to blend their collective online presence after a number of acquisitions. Coherence combined shopping carts of different web applications for a seamless online experience • Coherence enabled them to re-use their existing investment by sharing the HTTP session state and adding a single checkout application
Integration Web ApplicationsGlobal Airline • Application Migration • A major airline used Coherence to refactor and modularize their online applications • Using Coherence*Web to share HTTP session state their old application and new modular applications can be seamlessly integrated • Provided an incremental modernization route
Optimize Compute Processing Remove data bottlenecks for maximize compute processing • Fully utilize compute grids • Prevent computations stalling because of IO bottlenecks • Offload computation to Coherence, using it as a compute/data grid • Processing Incubator Pattern (used by OPIA)
Optimize Compute ProcessingMonsanto minimize data bottlenecks • Removing IO Bottlenecks • Monsanto wanted to accelerate and improve the computation of trial data for better and faster results. • Integrating Coherence made compute results available immediately, reduced database bottlenecks and enabled event driven processing • This eliminated ~30k man hours annually and reduced data turnaround time by 5 fold. It also improved data traceability, process flows and produced simpler and more standardized results
Optimize Compute ProcessingWackoviaminimize data bottlenecks Key Problems Customer Solution Key Benefits • Required absolute data availability for complex Grid Computations • Existing Compute Grid infrastructure suffering from data latency and throughput problems • Complex calculations so lengthy as to be outdated • Data Grid overlaid on Compute Grid • This enabled risk calculations to fully utilize the grid hardware by having real time access to in-memory data, improvingparallelization • Reduced critical risk computation from several days to under 1 hour!
Summary • Numerous customers are using Coherence to provide Internet/Cloud scalability to existing and new applications • Its unique combination of reliability and in-memory performance enables applications to provide the superior user experience customers now expect • The innovative real-time features of Coherence allow mobile “always online” applications to keep users of informed of changes as they occur
Join the Coherence Community http://coherence.oracle.com @OracleCoherence facebook.com/OracleCoherence blogs.oracle.com/OracleCoherence Oracle Coherence Users youtube.com/OracleCoherence coherence.oracle.com/display/CSIG Coherence Special Interest Group