Uploaded byAli Hodroj
E-Commerce and In-Memory Computing: Crossing the Scalability Chasm
The document discusses the challenges faced by today's e-commerce landscape, including the need for omni-channel personalization and the impact of holiday traffic on system performance. It highlights the role of in-memory computing as a solution to improve data processing speed and efficiency, particularly during peak traffic times. GigaSpaces, a provider of middleware software, offers in-memory data grids to enhance scalability and responsiveness for e-commerce applications.
More Related Content
Similar to E-Commerce and In-Memory Computing: Crossing the Scalability Chasm
E-Commerce and In-Memory Computing: Crossing the Scalability Chasm
- 1. E-Commerce and In-MemoryComputing Crossing the Scalability Chasm Ali Hodroj Director, Solution Architecture
- 2. 2 Agenda Introductionto GigaSpaces • Intro to GigaSpaces • Today’s e-commerce challenges Omni-Channel Personalization Holiday Traffic 1 2 3 Why In-Memory Computing? • In-Memory Data Grids E-Commerce Case Studies • GigaSpaces Use Cases for E-Commerce • Q&A
- 3. 3 About GigaSpaces GigaSpaces provides software middleware for deployment, management and scaling of mission-critical applications on cloud environments. GigaSpaces serves more than 500 large enterprises & ISVs, over 50 of which are Fortune-listed. 300+ Direct customers 40+ / 500+ Fortune / Organizations 75+ Cloud Customers 25+ ISVs
- 4.
- 5. 5 What canXAP do for you? Scaling the Data Tier Multi-site deployment & DR across remote sites Batch processing of large data sets Real time processing of large event stream Real time querying and analysis Online transaction processing of large datasets Scaling the Web Tier
- 6. 6 Today’s e-commercelandscape Is characterized by processing data from numerous sources Data Explosion Decision Time Compression Critical Time to Analytics Timeframe
- 7. 7 Latency Whyworry about it?
- 8. 8 • Numberof Holiday Season transactions grows exponentially • Tolerance for system response time reduces significantly Death by 120,000/sec Product lookups Source: Akamai PVS stats for a major US retailer
- 9. 9 Source: IBMCommerce Holiday retail readiness report 2014 An e-commerce site which takes more than 3 seconds to load will witness a 40% bounce rate
- 10. Omni-Channel: Seamlessly convergingaround one brand 10 across many channels
- 11. 11 Personalization Theshift from a “smart consumer” to “entitled consumer” Real-time Contextual
- 12. 12 Today’s modernretail ecosystem
- 13. 13 …and theOmni-Channel perspective is Consumer Experience Tier Infrastructure Tier
- 14. 14 But, theinfrastructure latency points are… Holiday Traffic Customer Segmentation Massive Data silos Analytics + Rules Synchronous bottlenecks And Cost Real-time retargeting 5x Traffic Generators Database bottlenecks
- 15. 15 In-Memory Computing Approach “In memory computing (IMC) … provides transformational opportunities. The execution of certain-types of hours-long batch processes can be squeezed into minutes or even seconds … Millions of events can be scanned in a matter of a few tens of millisecond to detect correlations and patterns pointing at emerging opportunities and threats "as things happen.”
- 16. 16 Data Grid Data Grid is a cluster of machines that work together to create a resilient shared data fabric for low-latency data access and extreme transaction processing
- 17. 17 Why In-MemoryComputing? Facebook keeps 80% of its data in Memory (Stanford research) RAM is 100-1000x faster than Disk (Random seek) Disk: 5 -10ms RAM: ~0.001msec
- 18. 18 E-commerce todaybelongs to the real-time world … Social User Tracking & Engagement Homeland Security Healthcare eCommerce Financial Services Internet of Things Real Time Search
- 19. 19 In-Memory ComputingUse Cases Omni-Channel Holiday Traffic Analytics + Personalization Content + Data Legacy Integration
- 20. 20 Context Omni-Channel Top 500 Retailer / Home Shopping TV - $8.6B in revenue Challenge Establish an omni-channel strategy by establishing a unified data layer that can control customer experience, sourcing, and supply chain Solution Medium scale XAP cluster (8 nodes) with built-in disaster recovery Results * Handles up to 250,000 Product reads/second * Unified user experience across TV, mobile, web, and operations Omni-Channel
- 21. 21 Terabyte-scale datagrid as a unified data and business logic infrastructure tier
- 22. 22 Context Content+ Data Top 500 Retailer / Home Shopping TV - $8.6B in revenue Challenge Globally available and fault tolerant shopping cart across data centers Solution Geographically distributed in-memory data grid with content replication Content + Data
- 23. 23 • E-commerce platform agnostic shopping cart storage tier • CDN-style caching at the application layer • Redundant shopping cart storage across multiple data centers
- 24. 24 Context HolidayTraffic Top 20 Retailer / 1,100 Stores - $19B in revenue Challenge Handle peak loads without over-provisioning for maximum traffic (following 2009 system crash resulting in loss of millions of $$) Solution Implemented inventory management on top of XAP within 4 months Results *Was N. America’s best performing e-commerce website on Black Friday 2010 Holiday Traffic
- 25. 25 • Holidayseason traffic spikes • Offers / Promotion campaign traffic • Elastic scaling
- 26. 26 Context Personalization Top 10 Retailer - $36B in revenue Challenge Adding social network functionality to e-commerce platform supporting over a million users every day –and up to 10x that on holidays Solution XAP powers Sears social networking search engine with in-memory storage and dynamic scalability Results * Fast effective social feature functionality, with performance to support peak usage Analytics + Personalization
- 27. 27 Real TimeBig Data Analytics Personalization platform using social networking data • Recommendation engines • Cross-Channel Analytics • Large-scale Clickstream analytics • Retrospective, event analytics • Behavior based analytics
- 28. 28 Context LegacySystems Integration Top 500 Retailer / Home Shopping TV - $8.6B in revenue Challenge Synchronous bottlenecks and MIPS cost pose a showstopper for omni-channel Solution Offload Mainframe data into a data grid, reducing cost Legacy Integration
- 29. 29 • 80%cost savings from mainframe access reduction • Integrate legacy data into modern omni-channel tier • Minimize contention with mainframe access
- 30. Check us out:www.gigaspaces.com Email us: [email protected] Call us: 646-421-2830 Follow us: @GigaSpaces, @CloudifySource