Cloud Computing

Cloud Computing

1. Overview

   • Cloud computing is a model which allows computing resources to be shared and configured across a network
   • Consumers can pay for what they use on a pay-as-you go model
   • Cloud services are accessible anywhere and anytime
   • Actors
     • Consumer
       • Access resources from centralised off site facility
       • Rent computer resources as needed
     • Provider
       • e.g.: Amazon web services
       • Offers computing resources to consumers on pay-per-usage model
       • Can offer storage, networking capabilities, server usage, and applications
       • Profit margins can be maintained by using low cost premises to store large scale data centers
     • Broker
       • Intermediary between provider and consumer
       • Assists consumer to select resources needed
       • Negotiates relationship between provider and consumer

2. Cloud Computing architecture

   • Cloud computing architecture can be divided into three tiers
     • Deployment models
       • Public
         • Infrastructure available to general public
         • Services offered on a pay per usage basis
         • Resources are scaled according to demand and without interaction
         • Resources maintained by third party
         • Lack of control over where resources are stored and with whom
         • Benefits small organisations by saving on capital expenses
       • Private
         • Available for one organisation
         • Run through enterprise data center
         • Managed by organisation or third party
         • In house data centers require significant capital expenses
         • Upfront costs more expensive than public cloud
         • greater control over resources and IT infrastructure
         • Fewer data security concerns due to better control over storage and access of data
         • Larger organisations have shown a preference for private cloud
       • Community
         • Supports two or more organisations with a similar interest
         • Popular among small to medium businesses
         • Can be managed internally or by third parties
         • Organisations agree on T&Cs to share resources
         • Must agree on security, privacy and compliance
         • Capital expenses shared among communtiy
         • Can implement higher data security and privacy than public clouds
         • Aspects of public and private
       • Hybrid
         • Most complex deployment model consisting of two or more cloud infrastructures
         • Combination of public and private allows for some resources to be managed in house and others to be outsourced
         • Outsource non-critical info and manage important info in house
         • Cost benefits of public cloud + control of data in private cloud
         • Need to consider security concerns of the public cloud component
         • Once secure can be more flexible than private and public cloud
     • Service models
       • IaaS
         • Amazon web services provides IaaS successfully
         • Service provider manages virtualisation, servers, storage and networking (see diagram)
         • advantages
           • Only pay for resources to deploy application
           • Traditionally good support, performance and security
           • Don't have to worry about complexities of hardware for deployment
       • PaaS
         • example of PaaS is the Google application engine
         • On top of IaaS the service provide manages the OS, Middleware and Runtime
         • Consumers provided with resources for developing, managing and deploying applications
         • advantages
           • Can use PaaS to develop, test and deploy applications
         • Billed for # of users and amount of resources consumed
         • Limited access to underlying infrastructure
       • SaaS
         • Salesforce.com is one of the most successful firms to use SaaS
         • Consumer uses service providers computer infrastructure and software
         • Multi-tenancy makes SaaS applications flexible and confiigurable
         • SaaS provider manages and controls all resources (compared with PaaS and IaaS)
         • Advantages
           • No complexity associated with infrastructure provision, maintaining software
       • Also HaaS, DaaS and BIaaS
     • Essential characteristics
       • On-demand self service
         • Provision of services on demand and without interaction
         • Real time response to change in demand
       • Broad network access
         • Access resources independent of location and through a variety of platforms
         • Promotes productivity through flexibility
       • Resource pooling
         • Pool resources to create a shared environment
         • Virtual and physical resources are pooled and redistributed to consumers
         • Can serve multiple customers with different needs
       • Rapid elasticity
         • Dynamically re-purpose resources to meet demand
         • Possible through virtualisation
         • Helps utilise resources effectively
         • Consumers can respond effectively to changes in demand
       • Measured service
         • Providers monitor, control and report on resources usage
         • Resources can be controlled and optimised through measuring
         • Measuring forms basis of pay-per-usage model

3. Benefits and Drivers

   • Cost Efficiency
     • Reduced upfront capital expenses
       • Transfer of risks and IT infrastructure to provider
     • Pay-per-usage model
       • Pay for what's used
   • Data recovery
     • Backups automatically generated
   • Minimal staff
     • Processes are automated
     • Staff only needed when problems can't be solved by automated processes
     • Cloud doesn't require staff for operations
   • Broad access
     • Can access resources from multiple platfoms in a range of locations
   • Scalability
     • Resources can be scaled elastically to meet demands
     • Doesn't require long term planning or commitments
   • Management
     • Providers responsible for maintenence, support and management
     • Updates and bug fixes handled by providers
     • Allows to focus on core business activities
   • Barriers to entry
     • Reduced cost and risk reduces barriers to entry
     • Small to medium businesses benefit the most
     • Resources can be scaled to meet demand

4. Risk and concerns

   • Security
     • Lack of control and visibility
     • Security relating to data transmission, application and 3rd party
     • Dependent on deployment
   • Privacy
     • Concerns due to the many devices, applications and customers on the network
     • Sharing of resources(multi tenancy)
     • Lack of visibility of enforcement of controls
   • Lock-in
     • Data lock in from change in ownership from consumer to provider
     • Providers has ownership and control over data
   • Interoperability and integration
     • Difficulty in transferring data between providers
     • Difficulty in integrating APIs with legacy systems
   • Availability
     • Absence of meaningful Service level agreements(SLA)
     • SLAs provide customers with an outline of the availability of servers, platforms and appl
     • If resources are not available it could have severe impacts on business
     • Risk SLA is not reliable or comprehensive or is deceptive

5. Management Issues

   • Costs and benefits
     • Other benefits
       • Fewer upgrades and maintenace costs
       • Reduced IT staff costs
     • Costs
       • Constant internet connection
   • Customer retention
     • reduced costs
     • On demand self service
   • Risk Management
     • Reduce risks through choice of deployment and service model
     • Investigate offerings from providers to ensure they have state of the art tech
     • Inspect SLA
       • Minimum levels for each service
       • remedies for failing to meet contract
       • Outline of infrastructure and security
       • Rights of consumers
     • People issues
       • Employee issues(resistance and training)
       • Provider issues(SLA, transparency and relationships)

6. Market Assessment

   • Vendors Amazon Web services - IaaS Microsoft IBM Oracle - BI PaaS
   • IaaS has the largest market share

7. Future trends

   • Hybrid cloud computing
   • Cloud centric design
     • Application designed for limitations and opportunities of a cloud model
   • Spending on private clouds

Imágenes y archivos del mapa