Allocated Load IT-capacities Forecast “Under-supply“ of capacities “Waste“ of Fixed cost of IT CAPACITY capacities IT-capacities Barrier forinnovations Actual Load TIME
Load Allocated IT Forecast capacities No “under-supply“ IT CAPACITY Reduction of Possible “over-supply“ reduction of IT-capacities in case of Reduction reduced load of initialinvestments Actual Load Time
Compute Compute Inactivity Period Average Usage Average Usage Time TimeOn & off workloads (e.g. batch job) Successful services needs to grow/scaleOver provisioned capacity is wasted Keeping up w/ growth is big IT challengeTime to market can be cumbersome Cannot provision hardware fast enoughCompute Compute Average Usage Average Usage Time TimeUnexpected/unplanned peak in demand Services with micro seasonality trendsSudden spike impacts performance Peaks due to periodic increased demandCan’t over provision for extreme cases IT complexity and wasted capacity
Packaged Infrastructure Platform Software (as a Service) (as a Service) Software (as a Service) Applications Applications Applications Applications You manage Data You manage Data Data Data Runtime Runtime Runtime Runtime Managed by vendor Middleware Middleware Middleware Middleware Managed by vendorYou manage O/S O/S O/S O/S Managed by vendor Virtualization Virtualization Virtualization Virtualization Servers Servers Servers Servers Storage Storage Storage Storage Networking Networking Networking Networking
A case study…
ProfileLargest dept of education in southern hemisphere
Problem before themTo improve year 8 scientifictesting, needed to run 65,000concurrent science examsacross 650 schools.Had to decide to on Cloud, vsself-hosted.
Self hosted Estimated up front cost of A$ 200,000To procure 300 server infrastructure for project
Self hosted Continuing costs of ?To maintain and store servers etc
Perspective.$250,000$200,000$150,000$100,000 $50,000 $0 Self hosting Microsoft Azure
A case study…
Case study in cost savings…
ProfileMajor supplier of data warehousing for hospitals…
Problem they solveWarehousing data for hospitals isexpensive, time consuming andfragmented across all the hospitals.
Self hosted Normal up front costs of € 100,000Per hospital to get high quality data recordssuch as updated physician list.
Case study in new services…
Love Clean Streetshttp://www.lewisham.gov.uk/doitonline/Pages/Report-it.aspx
love clean streets 87% improvement in clean up time responses£5.10 to take a 21% reductionclaim by phone in cases dropped to £1.10 on web
So what’s the path to the cloud?
Most of the artwork used in this presentation is licensed underCreative Commons by Frits Ahlefeldt, aka hikingartist Support his amazing craft at http://hikingartist.com