WEBINAR Big Data Fabric Drives Innovation And Growth
WEBINAR Big Data Fabric Drives Innovation And Growth Noel Yuhanna, Principal Analyst September 12, 2017. Call in at 10: 55 a. m. Eastern time © 2017 FORRESTER. REPRODUCTION PROHIBITED.
We work with business and technology leaders to develop customer-obsessed strategies that drive growth. © 2017 FORRESTER. REPRODUCTION PROHIBITED. 2
Moving to the next-generation of data architecture Self-service Digital business Real-time Lots of data BT Batch Limited data MIS IT Trend Automation © 2016 Forrester Research, Inc. Reproduction Prohibited.
Data architecture — moving to the future state From traditional state To current state To future state My data Integrated data Global integrated data Batch to semi-batch Right time Really real-time Many silos Integrated Everything is integrated Data centralized Data in the cloud Data lives everywhere IT-driven Departmental-driven Business-driven Technology-focused Customer-focused Business-focused Data management Customer data management Business data management Predefined state Basic level of self-service Self-service Bottom-up Some level of top-down Top-down Product-centric Customer-centric Business-centric Independent teams Collaboration Interdependent © 2017 FORRESTER. REPRODUCTION PROHIBITED. 4
Big data: trends › Adoption is growing rapidly — estimated at 55%. › About 30% of companies have failed on big data projects. › Mostly (85%) it’s IT, data scientists that are using the big data platform for now. . . but customers want to open the platform to more users/personas. › Most enterprises struggle with delivering an integrated view of information from big data, traditional data sources, and cloud. › Organizations are focusing on self-service big data strategy — customer analytics, 720 -degree view, and business analytics. › Real-time big data platforms are becoming more prominent. © 2017 FORRESTER. REPRODUCTION PROHIBITED. 5
Data lake: trends › Data-lake adoption currently is around 25% and likely to double by 2020. › Data lake is becoming critical for organizations to succeed — to deliver new insights and analytics to gain a competitive edge. › Data-lake failures — an estimated 25% of enterprises have failed to deliver on data lake largely because of budget, skills, or focus issues. › Data-lake technology is mature for supporting broad level of use cases. › Organizations are building multiple data lakes. › Security and governance are critical right from start. © 2017 FORRESTER. REPRODUCTION PROHIBITED. 6
Source: Big Data Fabric Drives Innovation And Growth Forrester report © 2017 FORRESTER. REPRODUCTION PROHIBITED. 7
Source: Big Data Fabric Drives Innovation And Growth Forrester report © 2017 FORRESTER. REPRODUCTION PROHIBITED. 8
Source: Big Data Fabric Drives Innovation And Growth Forrester report © 2017 FORRESTER. REPRODUCTION PROHIBITED. 9
Data lake architecture — Hadoop Streaming sources Consumption Clickstream Log stream Trans/oper. data Data discovery Hadoop Data preparation EDW OLTP Visualization BI/analytics External source Partners Saa. S © 2017 FORRESTER. REPRODUCTION PROHIBITED. EDW 10
Data lake architecture — moving into data lakes Streaming sources Clickstream Log stream Consumption Integration Security Classification Metadata Data discovery Data lake Trans/oper. data EDW OLTP Quality Processing Data preparation Enrichment Governance Visualization BI/analytics External source Partners Saa. S © 2017 FORRESTER. REPRODUCTION PROHIBITED. EDW 11
Big data fabric — glues lakes, EDW, and others together to drive enterprise data and analytics strategy Big data fabric Streaming sources Clickstream Log stream Trans/oper. data EDW OLTP Multiple data lakes Integration Security Classification Metadata Consumption Data lake Quality Processing Data discovery Data preparation Enrichment Predictive analytics Governance Systems of insight External source Partners Saa. S © 2017 FORRESTER. REPRODUCTION PROHIBITED. BI/analytics EDW AI/cognitive apps 12
What is a big data fabric? “Bringing together disparate big data sources automatically, intelligently, and securely and processing them in a big data platform technology, using data lakes, Hadoop, and Apache Spark to deliver a unified, trusted, and comprehensive view of customer and business data. . . ” Source: Forrester Research © 2017 FORRESTER. REPRODUCTION PROHIBITED. 13
Big data fabric architecture 2 “Bringing together disparate big data sources automatically, intelligently, and securely and processing them in a big data platform technology, using data lakes, Hadoop, and Apache Spark to deliver a unified, trusted, and comprehensive view of customer and business data. . . ” 5 4 3 1 © 2017 FORRESTER. REPRODUCTION PROHIBITED. 14
Big data fabric: trends › Big data fabric adoption currently is around 20% and likely to double over the next three years (2020). › Most organizations’ architectures evolve from DW — Hadoop — data lakes into big data fabric. › Data lake vendors are offering more automation and simplification with more solutions on the way — complete off-the-shelf are still evolving. › On an average, it takes between two months to over six months to build a big data fabric initial deployment. › Big data fabric in the cloud is growing rapidly. Estimated 20% of the all big data fabric deployment is in the cloud. . . © 2017 FORRESTER. REPRODUCTION PROHIBITED. 15
Top use cases for data lake/big data fabric › 360 -degree view of customer, product, and business › Fraud detection and risk analytics › Data landing/staging area for EDW, Hadoop, and data int. › Integrated analytics — across various silos › Various dashboard — customers, partners, etc. › Various vertical specific use cases. . . © 2017 FORRESTER. REPRODUCTION PROHIBITED. 16
Case study: Retailer leverages big data fabric/lake to deliver customer analytics › Background • Big data was spread across clickstream, social media, blog, several databases, logs, and data repositories. • Wanted integrated view across billing, revenue, and other customer data to better understand its customers and their usage patterns • Retailer also wanted real-time insights, immediate access to billed and unbilled revenue, and ability to upsell and cross-sell new products. › Solution • Retailer used a combination of Hadoop, streams, replication, Hive, and No. SQL to store, process, and access data from the data lake. • Some integration took place in Hadoop; others in-memory and Spark. • Plans to add more data sources — geolocation, customer preferences. . . © 2017 FORRESTER. REPRODUCTION PROHIBITED. 17
Case study: Financial services company uses big data fabric/lake to support fraud analysis › Background • With billions of events everyday, this large financial services company was facing a major challenge to detect, alert, and process fraudulent activities. • Data was spread across Oracle, SQL, Hadoop, Hive, files, streams. . . • Integrating data across these sources was a challenge, and with new sources being added, such as clickstream, web logs, and social media feeds, it had to look at a new approach. › Solution • Used data lake to store unstructured data, including logs and streams, and built models that integrated all relevant data sets in real time to accurately assess if any given activity was a fraud. • Unlike other banks and financial services companies that quite often had false positives, this financial services company was quite accurate in its analysis. © 2017 FORRESTER. REPRODUCTION PROHIBITED. 18
Case study: Manufacturing organization uses big data fabric/lake for Io. T analytics › Background • A large manufacturing company with hundreds and thousands of machinery and components and more than a dozen plants wanted a solution that could minimize machinery failures. • Some of the machine equipment was getting old, but the company wanted to ensure that replacements were being done for the right machines, parts, etc. › Solution • They installed sensors and additional devices to collect data that fed into the data lake/data fabric, along with other data sets. It streamed data to Hadoop in its data center, processed the data with historical data to determine machines likely to fail, wear out, and have parts issue. • Overall, the manufacturer claims to have eliminated many hours of machine outages every month and, thus, have related to savings of millions over the year. © 2017 FORRESTER. REPRODUCTION PROHIBITED. 19
Big data fabric — vendors Big Data Fabric Wave 2016 Last Forrester Wave vendors: Other vendors to look at: • Denodo Technologies • Cambridge Semantics • Global IDS • Cisco • IBM • Cloudera • Informatica • Hortonworks • Oracle • Pentaho • Paxata • Podium Data • SAP • Red Hat • Syncsort • SAS • Talend • Snowflake • Trifacta • Waterline Data Source: The Forrester Wave™: Big Data Fabric, Q 4 2016 Forrester report © 2017 FORRESTER. REPRODUCTION PROHIBITED. 20
Recommendations › Don’t boil the lake; start with a few data sources. › You don’t have to have a data lake to support a big data fabric architecture. › Create a big data/big data fabric team to ensure success. › Look at opportunities to upgrading from data virtualization architecture. › Leverage ML and AI — gradually. › Leverage cloud and hybrid. › Remember, big data fabric is an evolving architecture. . . © 2017 FORRESTER. REPRODUCTION PROHIBITED. 21
New: Forrester Insights for Android NOW AVAILABLE ON GOOGLE PLAY Download Forrester’s new Insights app for Android to: › Access research, insights and key takeaways to accelerate your projects and support your decision making. › Save reports and graphics to read on the device of your choice. › Receive notifications to stay abreast of the latest trends and insights relevant to your initiatives. Also available for i. OS. forrester. com/app © 2017 FORRESTER. REPRODUCTION PROHIBITED. 22
Noel Yuhanna nyuhanna@forrester. com Twitter: @nyuhanna Thank you FORRESTER. COM © 2017 FORRESTER. REPRODUCTION PROHIBITED.
- Slides: 23