The Impact of IoT on Data Centers
- 1. 1 EIT 4.1 Thinking Different: Data Centers and IoT Chris Crosby, CEO, Compass Datacenters
- 2. 2 Data Center World – Certified Vendor Neutral Each presenter is required to certify that their presentation will be vendor-neutral. As an attendee you have a right to enforce this policy of having no sales pitch within a session by alerting the speaker if you feel the session is not being presented in a vendor neutral fashion. If the issue continues to be a problem, please alert Data Center World staff after the session is complete.
- 3. 3 Accelerated Evolution • Rapid period of change • Old rules no longer apply • Structure • Roles • Decision making • Data center role shift • All-in-one versus matching need • Avoid load group mis-match • Scale matters
- 4. 4 Driving the Change • Convergence of data types • Large rich packets (video) • Billions of small packets (IOT) • Value of data • Inverse relationship with latency • End user requirements/expectations • Data must be as close to end user as possible – generational shift
- 5. 5 Not Ready for Primetime, Part 1 • The mega data center can’t do it all • Geography • Can’t get close enough to the customer • Latency becomes a serious issue • Applications processing capability • We used to call billions of tiny packets DOS • Now it’s business as usual • “Intelligence” isn’t distributed • Network • Fat pipes are no longer enough • Latency matters • Hetero v. Homogeneous • The big guys build for single apps • Too expensive/inefficient for most enterprises
- 6. 6 Not Ready for Primetime, Part 2 • Existing “network” structures • Single data center • 2-level (maybe) • Centralized with DR with usually some synchronous requirements • Centralized w/big pipe to regional • Regulatory environments pushing for more • Not optimized for “next generation” requirements • Difficult to hold “localized” content • Design didn’t anticipate volume • Ability to process volume in real time • Even if it isn’t your app set, it will affect the public network
- 7. 7 The Stratified Structure • Hierarchical structure • 3 levels • Data center(s) at each level have specific function • Based on division of labor • Improve processing capability • Distributed throughout the structure • Better support for converged architecture • Mission critical at all levels • Why? • $10M - $20M investment in just a few racks of gear • Even the “cloud” has 10-20 racks that run all the commodity infrastructure
- 8. 8 Cost of Redundancy Versus Software • Mission critical at the edge • Cost to put into the software versus the cost to put it into the hardware • Just like in your car, what you can afford today is different than 5 years ago Cost Time Data Center Software
- 9. 9 Roles Centralized hubs Primary applications processing/storage points Applications are “non-divisible” Edge data centers Regional centers Support one or more micro DC’s Perform regional processing/cache function Ex: Determine what goes upstream Micro data centers Initial interaction point Repositories for high demand content Located to provide lowest level of latency 9
- 10. 10 The Stratified Structure • Requirements • Flexibility • Must be able to adapt to shifts in demand • Dynamic • 80/20 rule • Security • At all levels • Ability to identify attacks • Dynamic re-routing of traffic • Mission critical at all levels • Maximize uptime • Protect high cost equipment • Converged equipment = more expensive • Tier III/IV (hub) • If size dictates, commodity compute storage at hub • Tier III (edge, micro) • Geography • Data centers located where they are needed • Especially at edge, micro levels • Location decisions driven by customer locations Commodity Level Converged Level
- 11. 11 Planning for Stratified • From tactical to strategic • More than we need a new data center in Cleveland • Need to think in 5-10 year periods • Not 12-24 months • Impact from outside, NOT just inside • Security • Network • Considerations • Applications • End users • Locations • Requirements • Expectations • Network • Technical and financial • Management and control
- 12. 12 Summary • Nature of data forcing the change • Majority of existing data centers/networks not ready • Stratified networks will become more prevalent • 2 major considerations: • Keep data closer to users • Lowest possible latency—generational shift in workers • Data center roles will become more specific • Based on hierarchical location • Planning will have to change • Tactical to strategic
- 13. 13 3 Key Things You Have Learned During this Session 1. Do we need bigger data centers 2. Where is the edge 3. What will we see in the next 24 months