Keith Inight, CTO at Atos - Software Defined Everything
6 likes1,493 views
The document discusses software defined networking (SDN) and its growth potential. SDN abstracts physical network infrastructure and exposes it through APIs to enable greater automation, policy-based orchestration, and reuse. It aims to increase agility and speed while reducing costs through more efficient use of commodity hardware. The document raises questions about how organizations are implementing SDN in areas like security, automation, availability, utilization, and standards.
Keith Inight, CTO at Atos - Software Defined Everything
- 1. Software Defined Everything A Business Perspective Image Credit: wavebreakmedia/Shutterstock
- 2. “The worldwide SDN market for the enterprise and cloud service provider segments will grow from $960 million in 2014 to over $8 billion by 2018” International Data Corporation, August 2014
- 3. • An architectural approach • Abstraction of the physical infrastructure • Services exposed via APIs • Enabling greater levels of automation, policy-based orchestration and reuse • Goals – Increasing the agility and speed of infrastructure through policy-based automation of processes. – Cost savings in capital expenditure by implementing resource pools containing commodity hardware. Source: Gartner (September 2014)
- 4. Security • Firewall rules centrally defined and implemented in all servers • Distributed not boundary
- 5. Agility • Enables end-to-end deployment of service without any physical changes • Capable of automation • Infrastructure implemented as incremental “pods” unrelated to project business cases
- 6. Automation • APIs provide the means to automate • But – Which APIs – Investment in processes, policies, templates, scripts… • Potential to reduce errors as well as reduce lead times
- 7. Where are you now? • Research • Evaluation • Deployment project • In production
- 8. Source: Gartner (September 2014)
- 9. Availability • Re-provision not repair on failure • Automation – reduction of human error – Really? • Complexity • Concentration of risk
- 10. Benefits – what do you believe are valid? • Agility • Security • Availability • Others….
- 11. Utilisation • Optimal placement of resources • Avoiding purchasing in advance of demand – Really? • How would you (or your contractor) purchase capacity? – Capital purchase of pods – Opex arrangement with supplier – Other?
- 12. Standards • Which will you follow? – VMware – OpenStack – Microsoft – Cisco – Other?
- 13. Recommended
- 14. Thanks for contributing Keith Inight CTO, Storage, DC, Applications and Tooling, Global Managed Services Office: +44 1773 550963 Mobile: +44 7733 312539 Email: keith.inight@atos.net
Editor's Notes
- #3: So whats all the fuss about Why should I be interested Whats in it for me…
- #4: Examples openflow VMware NSX – firewalls and load balancers become VMs OpenStack – open APIs
- #5: So – opportunity to have defence in depth and put firewalls between servers tiers etc. Rules are easier to manage – and disappear when last VM using them is terminated
- #6: Potentially minutes rather than weeks This isnt just about deploying a VM it is about all of the difficult stuff that normally takes time – firewalls, storage, ..
- #10: All eggs in one basket