Edge 2016 solving everyday problems with next generation mapping
- 1. © AKAMAI - EDGE 2016 Solving Everyday Problems with Next Generation Mapping Liz Borowsky, VP Platform Engineering, Akamai Technologies
- 2. © AKAMAI - EDGE 2016 What do I mean by mapping anyway?
- 3. © AKAMAI - EDGE 2016 A small scale example One Datacenter
- 4. © AKAMAI - EDGE 2016 Mapping: Load Balancing on steroids
- 5. © AKAMAI - EDGE 2016 Mapping Fundamentals Availability • Content is always accessible • Even when machines fail • Even when routing changes • Even when individual networks have problems Performance • The best Quality of Experience • Low latency • High throughput • Plenty of machine resources Primary Problem a CDN Mapping system wants to solve:
- 6. © AKAMAI - EDGE 2016 The two main approaches to mapping AnycastedNameserver Mapped 4.3.2.1
- 7. © AKAMAI - EDGE 2016 Anycast: Features • Conceptually simple • Easy to scale (to a point) Inherently somewhat centralized
- 8. © AKAMAI - EDGE 2016 Anycast: Availability BGP convergence time 30 seconds to up to 5 minutes* * https://labs.ripe.net/Members/vastur/the-shape-of-a-bgp-update
- 9. © AKAMAI - EDGE 2016 Anycast: Performance. US data, Gomez Last Mile tests
- 10. © AKAMAI - EDGE 2016 Anycast: Challenges – Lack of control • Load balancing done by BGP tuning • Hard to use resources efficiently • At the mercy of BGP churn and convergence • Scales only to a point
- 11. © AKAMAI - EDGE 2016 Nameserver Mapping: Availability Better Control 20 seconds to 1 min
- 12. © AKAMAI - EDGE 2016 Nameserver Mapping: Performance Control enables the ability to be widely distributed
- 13. © AKAMAI - EDGE 2016 Challenge: Nameservers far from end users Non-optimal mapping Poor performance
- 14. © AKAMAI - EDGE 2016 Great performance Solution: ECS (EDNS0-Client-Subnet DNS Extension*) *https://trac.tools.ietf.org/html/rfc7871
- 15. © AKAMAI - EDGE 2016 Solution: ECS (EDNS0-Client-Subnet DNS Extension) Average improvements from enablement with openDNS and GoogleDNS
- 16. © AKAMAI - EDGE 2016 Challenge: Scaling the internet Control = Opportunity
- 17. © AKAMAI - EDGE 2016 Challenge: Mobile End Users End users go through mobile gateways Combine clustering with ECS, but use the Gateway IP Subnet
- 18. © AKAMAI - EDGE 2016 Start of a solution: Mobile gateway clustering
- 19. © AKAMAI - EDGE 2016 Challenge: Scaling the internet Control = Opportunity
- 20. © AKAMAI - EDGE 2016 Challenge: Congested peering on the global internet Distribute load Avoid congestion
- 21. © AKAMAI - EDGE 2016 Distribute load away from congested peering • Measure throughput • Map away if throughput degrades • Use distributed platform to get content close to the end user • Offload overloaded peering links
- 22. © AKAMAI - EDGE 2016 Early results
- 23. © AKAMAI - EDGE 2016 Challenge: Congested peering on the global internet Distribute load Avoid congestion
- 24. © AKAMAI - EDGE 2016 The Future With control comes opportunity… • ECS in more networks • More metrics and data • More control to give better performance • More control to help the internet scale • Strategic anycast – best of both worlds
- 25. © AKAMAI - EDGE 2016 Questions? 問題 Fragen שאלות domande 質問 질문 frågor preguntas Вопросы pytania اﻷﺳﺋﻠﺔ vrae Sorular spørsmål