Best Practices for Building Successful Cloud Projects
Upcoming SlideShare
Loading in...5
×
 

Best Practices for Building Successful Cloud Projects

on

  • 3,873 views

 

Statistics

Views

Total Views
3,873
Views on SlideShare
3,772
Embed Views
101

Actions

Likes
3
Downloads
186
Comments
0

5 Embeds 101

http://www.gigaspaces.com 81
http://www.linkedin.com 10
http://www.slideshare.net 5
https://www.linkedin.com 3
http://46.20.118.12 2

Accessibility

Categories

Upload Details

Uploaded via as Microsoft PowerPoint

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment
  • Key Points:Focus on discussing the business aspects of Cloud Computing- What business issues are companies trying to solve with cloud computing?What value did moving to the cloud provide these companies?What concerns did these companies need to over come?
  • I came across this graphic a few months ago and thought it was applicable to undertaking a Cloud Projects and relating it to navigating through a mazeJust like any puzzle, there are multiple places to start, more than one path to success, pitfalls and dead ends along the way that you need to avoid.Making the wrong decision can at best case lead to lost time and at worse case lead to high costs and a failed projectAs you may have noticed…There is a lot of hype and buzz around cloud computingIt is cool to be cloud and every vendor now has “Cloud Ware”Remember when it was cool to be SOA? Well it’s not cool anymore Marketing teams around the globe are having a lot of fun coining new terms every dayAnd you need “Urban Dictionary” just to keep up with some of the latest meaningsI hope today to remove some of that confusion for you and help you navigate through the Cloud maze by providing you some real world experiences in what some of the customers that GigaSpaces is working with are thinking and doing with Cloud Computing.
  • I am not going to spend a lot of time on this slide but would like to level set on the terms that I will use today
  • First group is those projects that are looking for an “Enterprise Grade” IaaS/PaaS
  • The Perceived Risks of Newness of the Cloud Computing ModelWhile some kinds of software-as-a-service (SaaS)-style applications are well-established in mainstream use, use of APaaS is new. Therefore, most users are cautious, evaluating the viability of APaaS, especially considering that the leading software vendors have not yet endorsed this model of cloud computing.Proprietary Programming Models and Lack of StandardsThe model-based programming environments of most APaaS are nonstandard (and no standard for such programming presently exists). Other APaaS offerings that are attempting to use established programming languages and programming models for customizable multitenant computing are forced to reduce and alter the standard environments, losing the standard in the process. Thus, the current APaaS offerings lead to exclusive vendor lock-in, a serious barrier for mainstream adoption. A transition from the now-established programming models to the next generation of model-driven style of programming will require a strategic endorsement and backing from the leading software vendors.Data Security ConcernsUse of public cloud computing, whatever model is considered, for mainstream business applications is being delayed. This is because many organizations are concerned about the safety and security of their data. Thus, only the least-sensitive business applications are considered for cloud deployment. CRM is the most popular SaaS application category, and is less data-sensitive than other categories, because it does not address financial transactions. Some ERP and banking cloud applications are emerging, but have not yet reached the mainstream levels of adoption.Maturity of the Leading Software VendorsThe software industry giants have stayed away from the shared-everything, cloud multitenancy model. Instead, they are investing in the shared-hardware model, where elasticity occurs at the level of dynamically allocated virtual machines, while the platform stack above the virtual machine is the standard application server environment. Established software vendors and their users favor this approach, because it preserves their established business applications, tools and skills. Microsoft and other vendors likely will optimize the elasticity of the shared-hardware model to be competitive and, thus, will delay the emergence of a mainstream genuine shared-everything APaaS. However, pressured by the requirements of the SaaS market, the giants will be invested in delivering a genuine multitenant metadata-driven APaaS environment by 2012.Maturing of APaaS: Ready for Enterprise Computing Requirements?To be considered for enterprise-class systems, cloud application platforms must deliver reliable advanced levels of availability, manageability, security, data integrity, service-level agreement guarantees and functional completeness, including application integration, business process management, multichannel user experience, event processing, business intelligence, service-oriented architecture governance and other
  • This is truly a success story on how the economics of Cloud Computing enabled a companies business modelInitial problems occurred when Primatics started to develop their own platform by cobbling together various components:Unreliable, unstable and ended up causing system crashesEach crash required the risk analysis model to have to be restarted from the beginning causing them:Higher costs for re-use of cloud compute resourcesCustomer satisfaction due to missed SLA’sPrimatics learned that development practices that work in a static environment often break in a distributed cloud environment
  • Business drivers;Reduce the cost of delivery of their offering Expand their offering into un-tapped market segments that a large on-premise solution was too costly forWithout a large upfront investment they wanted to prove the market viability and the performance of a SaaS delivered call center application
  • AAR, a leading system integrator specializing in mission critical development projects, has delivered several turn-key solutions on top of GigaSpaces PaaS. GigaSpaces PaaS gives AAR a competitive edge, because it provides their customers the ability to launch new applications in a matter of weeks as opposed to months and at a fraction of the setup costs (development environment, testing environment, staging environment) and hardware costs of traditional projects. GigaSpaces PaaS allows integrators and solution providers to offer turn-key solutions on the cloud with unprecedented robustness, resilience, high availability and performance. Customers want cloud-based projects because they offer faster time to market, more flexibility and better scalability. Integrators benefit from dramatically reducing the total cost of projects, particularly the hardware and setup components, which can translate into improved profit margins or more competitive prices. In addition, integrators can generate ongoing revenues by offering to maintain the project on the cloud – GigaSpaces PaaS can even bill the customer automatically for these services using Amazon’s DevPay system.
  • Use less machine hours – with XAP you can reach 5-10 times the performance of standard implementations. This means you can deliver the same performance with less hardware. On the cloud, this translates into an immediate cost saving – by using less resources, you can save as much as 80-90% of the machine hours required per month.Use less bandwidth and storage – XAP uses its In-Memory Data Grid to keep your data in-memory, ensuring resiliency and consistency, which saves the need to purchase disk-based storage from the cloud vendor. XAP also runs your business logic collocated with your data, reducing the network bandwidth you’ll need to pay for.Run apps only when actually needed – XAP provides complete application lifecycle management, including automated deployment and cleanup (shutdown) of all application instances and middleware (load balancer, web container, database). This enables you to run applications and middleware only as long as it is actually needed. This results in substantial savings of machine hours on the cloud, especially in demo, testing and prototyping scenarios.No need to use the cloud for ongoing development – XAP lets you develop locally using an identical environment to the one running on the cloud. This means that ordinary development and testing does not require purchasing cloud machine hours.No need for additional cloud products such as Amazon’s map/reduce, parallel processing, automatic scaling and monitoring, SQS, SimpleDB, and loadbalancer, because XAP provides all these service out-of-the-box, included in its basic price.Test on the cloud and save in-house testing servers – XAP provides an identical environment for your applications both on and off the cloud, allowing you to test reliably on the cloud and then deploy applications in your local data center. Moving your testing environment to the cloud allows you to eliminate or re-purpose servers dedicated to testing, often more than 1/3 of the servers in a data center.Eliminate platform development costs – with most cloud solutions available today, you will need to develop some middleware/platform capabilities needed by enterprise applications. This is not only time-consuming and risky, but also requires special distributed computing skills. XAP saves you the time and trouble of building platform features yourself, by providing a comprehensive enterprise solution.

Best Practices for Building Successful Cloud Projects Best Practices for Building Successful Cloud Projects Presentation Transcript

  • John P. Giubileo
    EVP & GM, Americas
    GigaSpaces Technologies
    Best Practices for Building Successful Cloud Projects
  • Session Abstract: Where Business Meets the Cloud
    “Best Practices for Building Successful Cloud Projects"–
    A session providing best practices and guidelines for addressing common concerns while implementing cloud projects in today’s enterprise environments.
    Key Topics:
    What are some of the most common use cases today
    What are the most popular paths being used
    What are the most common concerns for adoption
    What is GigaSpaces doing in the cloud
    Real world examples of how our customers are leveraging the cloud
  • Navigating the Maze to a Successful Cloud Deployment
    Start Here
    Public Cloud
    Cloudburst
    Cloud Spanning
    SaaS
    Virtual Private Cloud (VPC)
    Cloudware
    Hybrid Cloud
    Private Cloud
    Cloudstorming
    External Cloud
    IaaS
    Vertical Cloud
    Cloud-Oriented Architecture (COA)
    Cloud Service Architecture (CSA)
    Cloudsourcing
    PaaS
    Inter-Cloud
    Cloud Portability
    Start Here
  • Standardizing on Terms for Today
    Software as a Service (SaaS)
    Off-premise delivery of an application
    Multi-tenant, Scalable
    Pay-per-Use or Pay-per-User subscription model
    Application Infrastructure as a Service (AIaaS)
    Servers as a Service with some application services
    Multi-tenant, elastic compute environment
    Pay-per-use pricing model
    Application Platform as a Service (APaaS)
    In-cloud application platform for the development and deployment of applications
    Multi-tenant, elastic, high-performance
    Similar middleware capabilities and components to on-premise application stacks
  • Business Drivers and Usage Scenarios
    IT as a Service AIaaS/ APaaS
    Increase application productivity and agility, focus on core business competency
    Looking for High-end enterprise-class capabilities, including scalability, performance, and reliability without in-house IT investments
    Reduce operating costs for new applications and business processes
    Gain a competitive edge
    Broader access than what is available from legacy on-premise applications
    POCs / Testing / Prototyping
    Better, Faster, and Cheaper way to keep up with the demands of the business
    Lower cost of sale/ Shorter Sales Cycle
    Faster Time to market
    A quick way to prototype and prove market and architecture viability
    Production scale testing
    SaaS enablement
    Alternative delivery model, market expansion, speed and cost to market
  • Today’s Popular Routes to the Cloud
    PaaS for Consumer Apps
    “Servers as a Service”
    Virtual Servers
    Virtual Servers
    IM Java/.Net App Server
    Virtual Servers
    Virtual Servers
    Virtual Servers
    Lightweight App Server
    In Memory Data Grid
    Non-Relational Database
    Cloud Computing Abstraction
    Non-Relational Database
    Infrastructure as a Service
    Infrastructure as a Service
    Infrastructure as a Service
    Simple but limited customization capabilities optimized for consumer applications
    Full control of the environment but
    does not eliminate the middleware complexity
    Flexibility of AWS + Simplicity of GAE + Enterprise Grade Capabilities
    PaaS for Enterprise Apps
  • Common Concerns for Cloud Adoption
    The Perceived Risks of Newness of the Cloud Computing Model
    Proprietary Programming Models and Lack of Standards
    Data Security Concerns
    Maturity of Leading Software Vendors Cloud Offerings
    Maturity of APaaS: Ready for Enterprise Computing Requirements?
  • Tips for Overcoming the Concerns
    The Perceived Risks of Newness of the Cloud Computing Model
    • Use mature solutions that are already offered and proven off the cloud
    • Take a gradual approach; e.g. first use the Cloud for testing, then for non-mission critical applications, etc.
    Proprietary Programming Models and Lack of Standards
    • Use standards based solutions that enable portability between on-premise and off-premise infrastructures to reduce vendor lock-in
    Data Security Concerns
    • Use the public Cloud for transaction handling, keep the database on-premise
    Maturity of Leading Software Vendors Cloud Offerings
    • They’re getting there …
    Maturity of APaaS: Ready for Enterprise Computing Requirements?
    • Some are! Let’s see how we and our clients did it.
  • Gartner: GigaSpaces IS Ready!
    GigaSpaces' eXtreme Application Platform 7.0 is one of the first industry examples of a cloud-enabled, extreme transaction processing platform.… Other vendors, including Oracle and IBM, plan to extend their DCPs with OSGi, Spring and Java EE elements. But none has yet architected a full cloud-enabled XTPP like GigaSpaces.


    Gartner Research Report, July 21st, 2009
  • Our Secret Sauce = Space Based Architecture (SBA)
    • A software architecture pattern for achieving linear scalability of stateful, high-performance applications
    Inspired by Yale’sTuple-Space Model
    Partitions the application and packages all middleware functions into a network of lightweight scalable units that live in a memory cloud.
    The best suited architecture for highly distributed applications & the Cloud
    Linear Scalability
  • GigaSpaces eXtreme Application Platform (XAP)
    An enterprise-grade application server for deploying and scaling Java and .NET applications under the most demanding and changing requirements.
    Single Integrated Application Platform
    Reduced complexity, no middleware integration, fewer license fees
    High Performance
    Less moving parts or network hops, guaranteed performance under any load
    Scalable on Demand
    Dynamic linear scalability, reduced cost of business growth, maximum hardware utilization
    Highly fault tolerant and resilient
    Zero downtime under unpredictable loads
    Open Architecture
    Minimize vendor lock-in, protect historic investments
  • A Complete Software-as-a-Service Platform
    Users
    Users
    Users
    • High-Performance
    • Auto Scaling
    • Multi-Tenancy
    • Single click deployment
    • Built-in portability
    • Cloud Mgt Framework
    • Support of Standards:
    • Development frameworks
    • Java, .NET, C++ and scripting languages
    Business Application as a Service
    Tenant App. View
    Tenant App. View
    Tenant App. View
    User Application
    Business Application
    Messaging
    Data
    Grid
    SLA
    GigaSpaces XAP
    Scale-out Application Server
    Platform as a Service (PaaS)
    Cloud Computing Framework for XAP
    Orchestration
    Monitoring
    Infrastructure as a Service
    CPU
    Storage
    Network
    Billing
    Public cloud
    Private Cloud
  • IT As A Service (Enabling the Business)
    Customer:
    Primatics Financial, offering a high-performance risk analytics/modeling solution on Amazon EC2 for investment banks and mortgage companies
    Cloud Business Drivers:
    On-Demand Performance: jobs must run in a matter of hours not days meeting the customer SLA’s at a low cost; application must auto-scale
    Pay-per-use model to alleviate the need for large upfront per customer infrastructure investments, improving company’s cash flow
    Primary Concern Addressed:
    Data Integrity concern addressed by keeping each clients data in a separate processing instance
    Results:
    Full system prototype in 3 weeks
    Performance of 60M records in 13 minutes = 77K records/sec on MySQL
    Auto-scaling application removing the need for human intervention
    Increased system reliability; reducing costs associated with processing restarts
  • SaaS Enablement
    Customer:
    A large global telecommunications service provider converting their on-premise IP telephony call-center application to a SaaS enabled global offering
    Cloud Business Drivers:
    Alternative delivery model of their legacy on-premise solution
    Expand solution into un-tapped market segments; primarily SMB
    Prove market viability and application scalability without large infrastructure investment
    Primary Concern Addressed:
    The risk of Cloud was minimized by using the same application platform and code base as their on-premise solution
    Results:
    Within a few weeks the application was deployed on a 150-node cluster on the cloud
    Response time of several milliseconds tested to 1 million concurrent users.
  • IT As A Service (APaaS)
    Customer:
    UK’s leading provider of mobile phone’s and broad band services.
    Cloud Business Drivers:
    Solution platform that reduced the time and cost to launch new services to market
    Required a real-time scalable front end to their consumer services offerings capable of handling unknown peak loads
    Primary Concern Addressed:
    Data security was ensured by persisting customer data to backend on-premise servers
    Results:
    In less than 2 months they created a platform that met their business requirements and increased their business agility
    Reduced their front end investment costs for new services
    Lowered their investment exposure for services that were rejected by the market
    Improved the performance and scalability of those services that had high adoption
  • Demos and Proof of Concepts
    Customer:
    GigaSpaces Technologies
    Cloud Business Drivers:
    Improve tools available (POC’s/demos) to the sales organization
    Enable customers to have a hands on experience to both GigaSpaces and Cloud Computing
    A need to reduce the cost of maintaining these environments
    Results:
    Fast proof of concepts that demonstrate the full functionality of our product both on or off premise.
    An easily maintained live demo environment available to the sales teams and customers anyplace in the world
    Ability to update infrastructure so latest capabilities are broadly available from a single place.
    Try it yourself for free: www.gigaspaces.com/mycloud
  • Final tips for utilizing the Cloud more efficiently
    • Use high performance platforms to reduce machine hours
    • deliver the same performance with less hardware => save as much as 80-90% of the machine hours required per month
    • Keep your data in-memory and use less storage
    • saves the need to purchase disk-based storage from the cloud vendor; alleviates data security concerns
    • Use solutions that are auto-scaling
    • Run apps only when actually needed to substantially save machine hours on the cloud
    • Don’t use the cloud for ongoing development
    • Use solutions that enable you to develop locally using an identical environment to the one running on the cloud => eliminate machine hours spent on ordinary development
    • Test on the cloud and save in-house testing servers
    • Use mature solutions to test on the cloud and then deploy applications in your local data center => eliminate or re-purpose servers dedicated to testing.
  • Q&AThank You!
    www.gigaspaces.com/mycloud