Cloud Computing Evolution Not Revolution Craig Magee Head of Global Infrastructure Strategy & Architecture ANZ Banking Group 26 November 2010
Cloud Computing Evolution Not Revolution Ontology Evolutionary Perspective Promises, Challenges and Remediation What we are doing at ANZ What next? 2
Cloud Computing - Ontology Definition: "The set of disciplines, technologies, and business models used to deliver IT capabilities (software, platforms, hardware) as an on-demand, scalable, elastic service" (Burton Group) Characteristics Shared Cloud Types Hybrid Service Based Private Community Public On Demand Elastic and Scalable Consumption-based Pricing Internet technologies Service Models Software as a Service (SaaS) Platform as a Service (PaaS) Pay-as-you-Go Pricing Models Ceiling Price Sources: Gartner (2010), Burton Group (2009), Tower Group (2010) Software Infrastructure as a Service (SIaaS) Hardware Infrastructure as a Service (SaaS) 3 Hybrid Private/Public Other (data transfer rate, user licences)
Evolutionary Perspective Adoption Diffusion of Innovation Structured Programming Object-Oriented Programming Java REST Internet technologies XML Component-based software engineering Fortran COBOL Simula Pascal LISP ARPANET Brad Cox: Software John McCarthy suggests: components computation may someday be organised as a public utility Mainframe C Smalltalk C++ HTML Mosaic browser MSFT COM CORBA J2EE SOAP SaaS IBM SOM Outsourcing ASP VMWare Client/Server Citrix SOA Computing Models Google AppEngine IBM Blue Cloud Amazon AWS MSFT Azure 1950 1960 1970 1980 1990 2000 2010 4
Promises, Challenges and Remediation Simplifies and optimises IT Defers and avoids costs Increases IT agility Enables faster ROI through better resource management Increases mobile workforce access to IT services Improves business continuity, inexpensive disaster recovery Additional expert IT management Poor or nonexistent SLAs (availability, performance, response time), difficulty of monitoring and trouble-shooting Cost visibility, hidden charges, different licencing and pricing models Regulatory requirements limit certain actions, e.g. data location Market immaturity, lack of interoperability, difficulty of migration risks vendor lock-in Increased vulnerability to external threats Inadequate Risk Management legal, compliance and reputation, Operational (data loss, privacy), complex value chain Vendor viability, providers that leverage other providers may be caught in the middle 5 Conduct due diligence, write performance management into the contract As with outsourcing, manage closely, decommission services not being used Understand what data must be where, ensure providers are able to comply Better standards, plan for alternatives, plan for migration Common security standards, partner collaboration Thorough due diligence, understand entire chain, 3 rd party verification, have alternatives Use standards where available, have alternatives, plan for migration
Regulator concerns about Cloud Computing and Outsourcing Regulators in every country in which we operate take a keen interest in our risk management and business continuity preparedness 6
What the ecosystem looks like IT Organisation Provider #1 Provider #2 Software Services Application Platforms Infrastructure Single vendor (SaaS) Service Interface Multivendor IT+Vendor (PaaS) IT+Vendor (SIaaS) Multivendor (IaaS) Multivendor 7
What we are doing at ANZ Virtualisation Large virtualisation landscape employing a policy of virtualise by default IT as a Service well understood and accepted in the bank Centralised and virtualised UNIX, AS400 and Mainframe Distributed Wintel environment 11000 O/S images, 5000 virtualised Significant investments in VMWare Vsphere Private Cloud Incorporating a service management wrapper around the current virtual environment to enable private cloud capabilities Centralising and consolidating servers and desktops into this cloud Ensuring we deliver to business demands of cost, agility, stability 5 4 Maturity Model Mission critical deployment - enterprise load balancing Non-mission critical deployment - enterprise Beyond x86 We see significant business and organisation value for cloud services across multiple platforms not all mission-critical workloads should run on Windows or single O/S platform Majority of FS firms here 3 2 Non-mission critical deployment - single line of business IT virtualisation and reference architecture Private not Public Security / Privacy / Customer Perception / Global Regulatons ANZ have sufficient scale to drive own private cloud Need different commercial relationships with partners and internal businesses 8 1 Pre-existence of shared-service usage as part of IT delivery or line-of-business operation Source: Gartner (2010)
What we think the future will look like The trend is for new technology to be implemented on top of existing technology Few systems are de-commissioned The result: Increased complexity + + =? IT will need to accommodate composites (multiple generations) of technology in the same landscape We see emerging simplification patterns, such as integrated systems (e.g. Oracle Exalogic Elastic Cloud) to address the complexity challenges 9
Thank you Craig Magee Head of Global Infrastructure Strategy & Architecture ANZ Banking Group craig.magee@anz.com 10