Machine Learning for Your Enterprise: Operations and Security for Mainframe Enterprises
0 likes415 views
The document outlines a webcast focused on machine learning applications for mainframe operations and security, detailing the challenges and use cases for enterprises. It includes information on machine learning concepts, polling questions regarding big data platforms, and features of a machine learning toolkit. Participants can submit questions during the presentation, and resources will be made available after the session.
Machine Learning for Your Enterprise: Operations and Security for Mainframe Enterprises
- 1. Machine Learning for Your Enterprise: Operations and Security for Mainframe Enterprises
- 2. Housekeeping Webcast Audio: – Today’s webcast audio is streamed through your computer speakers. – If you need technical assistance with the web interface or audio, please reach out to us using the chat window. Questions Welcome: – Submit your questions at any time during the presentation using the chat window. – We will answer them during our Q&A session following the presentations. Recording and Slides: – This webcast is being recorded. You will receive an email following the webcast with a link to download both the recording and the slides. 2
- 3. Session Abstract and Speakers Machine Learning for Your Enterprise: Operations and Security for Mainframe Enterprises – What is Machine Learning: The Vision vs. Reality – The Challenges Driving Automated Mainframe Operations – Use Cases for Machine Learning at Mainframe Enterprises The presenters will also do an open Q&A with you and discuss results from our interactive quick- polls conducted during the session. 3 Syncsort Confidential and Proprietary - do not copy or distribute Zhe “Maggie” Li Chief Architect Steven Menges, Director, Product Management David Hodgson, General Manager/CPO
- 4. Speakers 4 Syncsort Confidential and Proprietary - do not copy or distribute Zhe “Maggie” Li Chief Architect
- 5. Speakers 5 Syncsort Confidential and Proprietary - do not copy or distribute David Hodgson, General Manager/CPO
- 6. Machine Learning Poll #1 Syncsort Confidential and Proprietary - do not copy or distribute 6 Q1.Which Big Data analytics platforms does your company use today? o Hadoop o Splunk o Elastic / ELK stack o SAS o Other Data Warehouse o Don’t Know (Check all that apply)
- 7. 77Syncsort Confidential and Proprietary - do not copy or distribute Enterprise Computing – Mainframe?
- 8. 88Syncsort Confidential and Proprietary - do not copy or distribute 2000+ Organizations Overall 71% Fortune 500 2.5 BillionBus. Transactions / day / per MF 23of Top 25 US Retailers of World’s Top Insurers10Top World Banks92 Source: IBM Mainframe in Enterprises Today
- 9. Enterprises With Mainframes Facing New Challenges Security – Mainframes are connected to mobile, IOT, cloud and open systems – External attacks – Internal threats (unknown unknown) Automation of IT Operations – Transactions grow exponentially – Increased complexity – Aging problem for mainframe skilled population – Lower costs required
- 10. Machine Learning for the Enterprise - No Longer a “Future?” Syncsort Confidential and Proprietary - do not copy or distribute 10
- 11. What is Machine Learning? “Machine Learning is a fascinating field of artificial intelligence research and practice where we investigate how computer agents can improve their perception, cognition, and action with experience. Machine Learning is about machines improving from data, knowledge, experience, and interaction…”
- 12. Machine Learning Machine learning uses algorithms to build analytical models and help computers “learn” from data. It makes predictions and uncovers hidden insights about relationships and trends.
- 13. Machine Learning
- 14. Categories of Techniques Supervised Learning Unsupervised Learning
- 15. Categories of Techniques Supervised Learning: Have the idea that there is a relationship between the input and the output. • Regression model: predict continuous valued output • Housing price • Weather forecast • Classification model: map input variables into discrete categories. • Identify cancer • Handwriting detection Unsupervised Learning: little or no idea what our results should look like. • Clustering: • Market segmentation • Social network analysis • Anomaly detection
- 16. Predict with Machine Learning actual data input y = H (X) prediction = H (X) hypothesis new input
- 17. The Vision vs. Reality
- 19. Machine Learning Poll #2 Syncsort Confidential and Proprietary - do not copy or distribute 19 Q2. Is Mainframe SMF and/or “log” data going into your big data platform/repository? o Yes, it is being streamed into it today o Yes, it goes into it via periodic batch/other input method o No, but that data has been requested/is desired o No o Don’t Know
- 20. Reminder 20 Syncsort Confidential and Proprietary - do not copy or distribute Type in your questions at any time during the presentation using the chat window. We will answer them during our Q&A session following the presentations or afterward.
- 21. Examples 21 Syncsort Confidential and Proprietary - do not copy or distribute
- 22. Critical Machine Data Streamed to a Big Data Platform
- 23. Critical Mainframe Machine Data Normalized and Streamed to Splunk with Ironstream® Log4jFile Load SYSLOG SYSLOGD logs security SMF 50+ types RMF Up to 50,000 values DB2SYSOUT Live/Stored SPOOL Data Alerts Network Components Ironstream API Application Data Assembler C COBOL REXX USS
- 24. Machine Data Machine Learning Platform - High Level Architecture Send TCP Send HTTP Send Kafka Predictive Analytics With Machine Learning Splunk/ Hadoop/ Cloud Get TCP Get HTTP Consume Kafka Automation tools Other Apps Operator commands Dynamic reconfiguration Data collection Data Transformation Data lineage/Metering data feedback z/OS Ironstream Configuration GUI
- 25. Splunk Platform Machine Learning Toolkit The Machine Learning Toolkit App delivers new SPL commands, custom visualizations, assistants, and examples to explore a variety of ml concepts. Assistants: – Predict Numeric Fields (Linear Regression): e.g. predict median house values. – Predict Categorical Fields (Logistic Regression): e.g. predict customer churn. – Detect Numeric Outliers (distribution statistics): e.g. detect outliers in IT Ops data. – Detect Categorical Outliers (probabilistic measures): e.g. detect outliers in diabetes patient records. – Forecast Time Series: e.g. forecast data center growth and capacity planning. – Cluster Numeric Events: e.g. Cluster Hard Drives by SMART Metrics
- 26. The Basic Process of Machine Learning Clean and transform your data – To meet the analytics explicit requirements Fit the model – Toolkit features 27 algorithms for fitting models – Over 300 open source Python algorithms in the add-on Validate the model – Each assistant provides a few methods in the validate section Refine the model – Adjust the parameters to improve the metrics Deploy the model – Deployment actions fall into the following categories • Make prediction or forecast • Detect outliers and anomalies • Trigger or inform an action
- 27. Splunk Platform Machine Learning Visualizations
- 28. 2828 Use Case Areas Syncsort Confidential and Proprietary - do not copy or distribute • RACF/ACF2/TSS Authentications • TSO account & login activity • FTP sessions & file activity • Sensitive data access & movement (PII/PHI) • Configuration settings (e.g. FISMA) • IRS Pub 1075 • Incident triage • Response times/SLAs • Latencies • Exceptions • Resource utilization • Anomalous behavior detection • Glass table view of entire service • Predictive analytics Security Trouble- Shooting Health Monitoring Compliance
- 29. Summary 29 Syncsort Confidential and Proprietary - do not copy or distribute
- 30. Questions and More Information Additional Questions for David and Maggie? For More Information: syncsort.com/ironstream blog.syncsort.com/ Try Ironstream for Free: syncsort.com/ironstreamstarteredition Comments/Other: Steven Menges: smenges@syncsort.com 30 Syncsort Confidential and Proprietary - do not copy or distribute