Veritas + MongoDB
Download as PPTX, PDF0 likes1,273 views
This document summarizes Veritas' experience moving their MongoDB deployment from an on-premise Enterprise edition to MongoDB Atlas on the cloud. Some key points: - Veritas migrated to improve scalability, simplify administration through Atlas self-service tools, and move to an OpEx model. - The migration process took only 10 minutes to correct some initial issues and was completed with minimal downtime. - Moving to Atlas reduced administrative time by around 90% and simplified management tasks like creating new environments. - Performance monitoring and optimization tools in Atlas helped Veritas identify query improvements. - Overall, Atlas solved Veritas' challenges around licensing, staffing, and infrastructure management while improving performance.
Veritas + MongoDB
- 1. Veritas + MongoDB JP Aubineau, Sr. Principal Product Manager Ben Baker, Principal DevOps Engineer
- 2. Agenda AboutVeritas MongoDB +Veritas Appliances How we saved 90% of operational costs moving to Atlas Copyright © 2018 Veritas Technologies2
- 3. Digital Compliance - Information Map - Enterprise Vault - Data Insight - eDiscovery Platform Software-Defined Storage * - Block: InfoScale - File: Access - Object: Cognitive Object Storage Data Protection * - Enterprise: NetBackup - SMB: Backup Exec - Cloud: CloudPoint | SaaS Backup IT Resiliency / Disaster Recovery - Resiliency Platform * Indicates market leadership Shared Intelligence - Data Visibility - Data Classification - Predictive Insights Our Mission To enable people to harness the power of information Copyright © 2018 Veritas Technologies3
- 4. Copyright © 2018 Veritas Technologies4 Our Pedigree Gartner Market Share: “Enterprise Infrastructure Software, Worldwide, 2017” May 2018 #1Backup and Recovery Software Market Share #1Software Defined Storage Management Market Share #1Integrated Backup Appliance Market Share 19Years Of Data Protection Magic Quadrant Leadership Recognized Leadership IDC: “PBBA Market Tracker Q1 2018”
- 5. Copyright © 2018 Veritas Technologies5 MongoDB + Veritas Appliances
- 6. • Established in 2010 • >25,000 systems deployed • Primary use cases – NetBackup (Backup & Recovery): 52xx/53xx – Access (Long Term Retention): 33xx • Scales from 4TiB to 2PiB • Offers simplification & lower TCO – Deployment – Performance – Maintenance Copyright © 2018 Veritas Technologies6 Veritas Appliance Business
- 7. AutoSupport Client • HW Monitoring • System Inventory • Alerting • Diagnostics collection • Capacity Utilization Internet • 24 x 7 Appliance Monitoring • Automated HW fault analysis • Proactive outreach viaCall Home Services • Field Service Dispatch Call Home Services Field Service Appliance PortalCustomer Machine Learning Use Case: Veritas AutoSupport Copyright © 2018 Veritas Technologies7
- 8. • Major transformation in our telemetry platform in 2014 • Architects preferred a document-based DB over relational Key/Value DB – Flexibility in Schema – Start small & grow – Favorable licensing format for embedded platform use • Embedded community version on premise • Enterprise / Atlas for cloud Copyright © 2018 Veritas Technologies8 MongoDB is core to our Appliance Business
- 9. Use Case: Veritas Predictive Insights Copyright © 2018 Veritas Technologies9 AI Built on 3 Years of Veritas AutoSupport Telemetry – 100’s of millions of telemetry data points – Leverages 100’s of man-years of Veritas Support Serviced data Continual Self-Learning – AI and ML utilizing 8 unique Veritas computational models – Constantly improves insights and accuracy Immediate Value – Eliminates alert fatigue – Proactive support enablement – Increases ROI by reducing downtime
- 10. Copyright © 2018 Veritas Technologies10 Challenges Staff Attrition “Manual” Automation Poor Reporting Procurement Difficulty 4 years post-transformation License was up for renewal this year Needed to move up from 3.2 to a supported versionAlso:
- 11. Copyright © 2018 Veritas Technologies11 MongoDB Atlas: Success Story Ben Baker
- 12. Intent to scale up aggressively in telemetry expansion across product portfolio – Volume of data expected to grow rapidly Optimize infrastructure for speed and agility – Design for ML/AI & deep analytics – Improve reporting capabilities Reduce administrative & licensing procurement burdens – Move from CapEx to OpEx model – Enable self-service administration for Engineering Increased resiliency & automation – Improve BCDR process Copyright © 2018 Veritas Technologies12 From Enterprise to Atlas: Goals
- 13. • Cost comparison was a wash but promise of administrative reduction was significant • Atlas licensing “pay as you go” was immediately favorable to address our scaling needs • Private instance & VPC peering was big plus for our security requirements • Migration tools and upgrade-in-place options helped simplify planning Copyright © 2018 Veritas Technologies13 From Enterprise to Atlas: Decision Process AtlasEnterprise
- 14. Copyright © 2018 Veritas Technologies14 From Enterprise to Atlas: Execution Strategy Existing MongoDB 3.2 Replica Set App1 App1 App1 App2 App1 App3 App1 App4 App1 App5 App1 App6 Connection to existing MongoDB 3.2 (Prior to migration) VPC Peering Connection to MongoDB Atlas 3.4 (Post Migration) Setup Atlas 3.4 Cluster1 3 Take snapshots of all EC2 instances 4Suspend traffic Load Balancer 7 Enable traffic 5 Migrate apps to new Atlas DB 6 Perform internal testing 8 Perform E2E Testing 2 MongoMirror Server migrates data to Atlas 9 Decommission Old 3.2 Instances
- 15. • Compass connectivity issues • Network CIDR clarity – Needed to connect from single Compass server to multiple Atlas projects • Didn’t realize each Atlas project required a separate network block – Recreated new projects with new CIDR blocks, migrated data • Incredibly easy to complete ~ 10min of administrative time overall to correct Copyright © 2018 Veritas Technologies15 New Challenges since moving to Atlas
- 16. • Overall downtime impact to Appliance Customers: <5 minutes – Zero support cases / escalations – Zero P1/P2 internal issues post-release Copyright © 2018 Veritas Technologies16 Results! • Subsequent administrative time reduced by roughly ~90% – Spin up of new dev/prod replica environments within minutes • (compared to hours in past) – Atlas Performance Advisor helped discover new index opportunities • Noted improvement in query efficiency and application response
- 17. • Memory Pressure • I/O Bottlenecks • Query execution • Primary/Secondary Member evaluation • Simple customizable interface Copyright © 2018 Veritas Technologies17 Mongo Metrics simplifies performance issue discovery
- 18. • Solved our key challenges – Licensing, Staffing, Reporting • Eliminated infrastructure challenges – No longer have to worry about the platform • Simplified management & remedial tasks • Next Steps / Explorations – Consolidation of ArangoDB data into Atlas – Using BI Connector for reporting integration Copyright © 2018 Veritas Technologies18 In Summary
- 19. Thank you! Copyright © 2017 Veritas Technologies. All rights reserved. Veritas and the Veritas Logo are trademarks or registered trademarksof Veritas Technologies or its affiliates in the U.S. and other countries. Other names may be trademarks of their respective owners. This document is provided for informational purposes only and is not intended as advertising. All warranties relating to the information in this document, either express or implied, are disclaimed to the maximum extent allowed by law. The information in this document is subject to change without notice. JP Aubineau & Ben Baker Copyright © 2018VeritasTechnologies LLC
Editor's Notes
- #3: Veritas spends more on Research and Development than all of our competitors combined.
- #4: Veritas spends more on Research and Development than all of our competitors combined.
- #5: - We are the leading data management company. - 86% of Fortune 500 rely on Veritas
- #7: Appliances mimic what Atlas does for Mongo! Replaced a flat file configuration structure and embedded Sybase DB with single Mongo instance Key Goals Avoid cloud vendor lock-in High performance & elastic scalability Reliable support model for mission critical systems Minimize development change/process from client side infrastructure
- #9: Replaced a flat file configuration structure and embedded Sybase DB with single Mongo instance Key Goals Avoid cloud vendor lock-in High performance & elastic scalability Reliable support model for mission critical systems Minimize development change/process from client side infrastructure
- #11: Infrastructure management resource attrition – DevOps resources were shrinking Poor visibility into DB metrics & performance Difficult internal procurement process Internal procurement process limited our ability to prototype and expand on-demand with Enterprise License model Ensuring license compliancy Procurement bottlenecks
- #12: Running on 3.2, which was going EOS this year, and needed to upgrade
- #15: Planning took ~3 months Migrating from 3.2.19 to 3.4 Application upgrades were required MongoDB Professional Services used as a guide for developing runbook Application cutover 3 Mission Critical applications, 5 Auxiliary applications Compatibility concerns Upgrade of MongoDB Drivers required Stricter validation of index specifications Approach for 3 environments Development Pre-production Production
- #17: Talk more about other new metrics