![Conceptual model
Opera3onal intelligence applica3ons
Analy3cs applica3ons & open APIs [output pub/sub pipeline]
Streaming
telemetry
State data NMS/EMS Polling data
Ÿ Ÿ Ÿ
Ÿ Ÿ Ÿ
Data aggrega3on & open APIs [input pub/sub pipeline]
Disk storage
Consume
Ingest
Collect
Transform, fuse & enrich [ETL]
Extract
Specialized tools & general BI tools
Business applica3ons User interface Present
Process
Dashboards & reports
In-memory
storage
Analy3cs engines](https://image.slidesharecdn.com/onugspring2017frameworkstephencollins20170422-compressed-170619185548/85/A-Framework-for-Infrastructure-Visibility-Analytics-Operational-Intelligence-14-320.jpg)
More Related Content
What's hot (20)
Similar to A Framework for Infrastructure Visibility, Analytics & Operational Intelligence (20)
Recently uploaded (20)
A Framework for Infrastructure Visibility, Analytics & Operational Intelligence
- 2. A Framework for Infrastructure Visibility, Analytics & Operational Intelligence Stephen Collins Principal Analyst, ACG Research
- 4. Visibility Analy3cs Infrastructure monitoring Root cause analysis Network monitoring Op3miza3on Traffic engineering Intelligence Automa3on Digital QoE Opera&onal intelligence use cases Value / Return Cost / Complexity Sweet spot Bleeding edge Low-hanging fruit Leading edge Security DDoS protec3on Service assurance
- 6. Vic&ms of our own success Increasing scale Nonstop growth in users, devices, end points Prolifera3on of applica3ons and services Global enterprise and service provider nets Web-scale enterprises on the Internet Hyperscale data centers Increasing complexity User/device diversity & heterogeneous networks Increasing number of security threats Cloud-based applica3ons and services So5ware-defined, virtualized infrastructure IoT adds another layer of complexity 1970 2020 1980 1990 2000 2010 Terminals Mainframes Minicomputers PCs Worksta3ons LANs Servers WANs Internet Web 2.0 Mobile Apps XaaS Cloud-na3ve Web-scale
- 8. Opera&onal silos IT opera3ons Security opera3ons Customer Service Network opera3ons DevOps Compliance Opera3onal applica3ons & tools Analy3cs Monitoring Control Config Opera3onal applica3ons & tools Analy3cs Monitoring Control Config Opera3onal applica3ons & tools Analy3cs Monitoring Control Config Opera3onal applica3ons & tools Analy3cs Monitoring Control Config Opera3onal applica3ons & tools Analy3cs Monitoring Control Config Ÿ Ÿ Ÿ Duplicate data & func3ons, inability to share data, high costs, proprietary … Ÿ Ÿ Ÿ
- 9. Big Data is the game changer Huge volume of data Distributed processing model High velocity of data Real-3me streaming analy3cs Wide variety of data Structured & semi-structured Flexibility & performance Schema-on-read analy3cs Rich set of open APIs RESTful, publish/subscribe pipelines Machine learning & AI Predic3ve analy3cs Thriving open source community
- 11. Open visibility & analy&cs framework Applica3ons & tools Big Data analy3cs & open APIs Applica3ons & tools Applica3ons & tools Applica3ons & tools Visibility source Visibility source Visibility source Visibility source Ÿ Ÿ Ÿ Common analy3cs plaforms, shared data, less duplica3on, lower costs, open … Network operaFons Security operaFons DevOps IT operaFons Ÿ Ÿ Ÿ Visibility source
- 13. Analy&cs-driven opera&onal & business intelligence Streaming telemetry Fuse, enrich & transform Real-3me analy3cs Batch analy3cs Opera3onal intelligence Business intelligence Big Data IT opera3ons Security opera3ons Customer Service Network opera3ons DevOps IT Planning CXO Marke3ng Compliance Business opera3ons Data at rest Time scale: seconds / minutes Time scale: hours / days Analyze first, then store Opera3onal func3ons Business func3ons
- 14. Conceptual model Opera3onal intelligence applica3ons Analy3cs applica3ons & open APIs [output pub/sub pipeline] Streaming telemetry State data NMS/EMS Polling data Ÿ Ÿ Ÿ Ÿ Ÿ Ÿ Data aggrega3on & open APIs [input pub/sub pipeline] Disk storage Consume Ingest Collect Transform, fuse & enrich [ETL] Extract Specialized tools & general BI tools Business applica3ons User interface Present Process Dashboards & reports In-memory storage Analy3cs engines
- 15. Pervasive visibility – instrument everywhere Enterprise networks Data centers & WANs Service provider networks Provide visibility to enterprise customers Public & private cloud infrastructure Physical server, OS, hypervisor, containers Internet infrastructure Topology, DNS, CDNs End points Local, remote & mobile ApplicaFons Correlate applica3on behavior with infrastructure
- 16. Pervasive visibility – instrument everything Streaming telemetry Feeds real-3me analy3cs Fine-grain, mulF-user controls Enable/disable visibility on-demand SoLware instrumentaFon Mul3ple layers – physical & virtual Hardware instrumentaFon Network processors & ASICs SoLware probes and agents Both passive and synthe3c monitoring
- 17. Pervasive visibility – open data models • Network & elements • Probes & NPBs • Cloud infrastructure • Apps & end points Many data sources • Packet & flow data • KPIs, network state • Syslog, server state • App & device data Many data types • Data cleansing, filtering & processing • Open data formats • Modeling languages Open data normaliza3on AggregaFon InstrumentaFon
- 18. Big Data trends - technology Publish/subscribe pipelines Both data inges3on & extrac3on Data normalizaFon / ETL Major source of complexity Open data formats & schemas No one-size-fits-all data repository In-memory, flash-based, disk-based HDFS, NoSQL, key-value, column-oriented No one-size-fits all analyFcs engine Hadoop MapReduce batch processing Spark & Storm streaming analy3cs Support for SQL queries Machine learning engines
- 19. Big Data trends - business Industry-wide Big Data skills gap Vendors, service providers & enterprises Complex value chain High degree of custom integra3on Big Data is a moving target Technology and product risk Need for open, turnkey soluFons Vendor-sourced, managed services or SaaS-based Cloud-naFve implementaFons (public or private) Elas3c, hyper-scalable, highly resilient Both private cloud & public cloud