From infrastructure to applications Where cloud computing is at and where it’s headed
Where it started In 2006, Amazon Web Services introduces Elastic Compute Cloud (EC2) and Simple Storage Service (S3). It showed what true on-demand computing could look like. No more waiting in lines for resources. No more overprovisioning.
How it grew SaaSIaaS PaaS Private/enterprise cloud
Where it’s at: SaaS Big business even for small businesses Services for everything Open APIs make platforms It’s for real
Where it’s at: IaaS Battle for supremacy Amazon Web Services Microsoft VMware Google Rackspace Balancing economies of scale and substance Security Performance Control Features
Where it’s at: PaaS The future Developer-focused (no servers; just push code) Service-oriented Little money in next-generation apps Google and Microsoft added IaaS Choices, choices … Heroku, Google App Engine, Microsoft Windows Azure, DotCloud, CloudBees, Cloud Foundry, AppFog, Engine Yard, Red Hat Openshift, Apcera
Where it’s at: Private/enterprise cloud Service providers are buying cloud software Are traditional businesses? Downloads vs. paying customers In a state of flux Is the goal to mimic AWS, or to operate like AWS? Neither? OpenStack? AWS? CloudStack? Hybrid clouds? Isn’t it just virtualization management? VMware/EMC spin out Pivotal Initiative
Biggest drivers to date (public cloud) Cost On-demand pricing Subscription vs. license Flexibility Scale/geography Productivity Time to market
Biggest drivers going forward CollaborationMobileBig Data
Collaboration (aka “social”) Inter-app More SaaS APIs mean more sharing data among apps Intra-company Sharing data Project management Tracky Communication Inter-company Cisco’s big data intermediaries Genomics/pharmaceuticals Collaborate anywhere Device Network
DNAnexus
Mobile Challenges of a mobile-first world Latency Akamai Online/offline synchronization Couchbase Processing power OnLive Developers!!! This is where PaaS becomes critical lightweight development/ops Heroku, Flurry, Parse, Keen.io
Parse
Big data Applications and infrastructure (e.g., Metamarkets and Elastic MapReduce) Optimal architecture either way More native web/cloud data Collective intelligence Malware, systems management, etc. Scalability Up and down
BloomReach
About those private clouds … Yes, deliver services like cloud providers, but … ALSO OPERATE LIKE CLOUD PROVIDERS.
How to operate like Google Cheap, efficient hardware Open Compute Project Dell C Series Software-defined everything Networks Storage Data centers Smart co-location Efficiency Proximity Modularity
Asking the right question In the long term: Ask not how you can move your current applications to the cloud, but what you can do differently because of the cloud.