Df10200 gartner

2,597 views

Published on

Published in: Technology, Business
0 Comments
3 Likes
Statistics
Notes
  • Be the first to comment

No Downloads
Views
Total views
2,597
On SlideShare
0
From Embeds
0
Number of Embeds
4
Actions
Shares
0
Downloads
244
Comments
0
Likes
3
Embeds 0
No embeds

No notes for slide
  • The hype around cloud computing grew during 2009. Vendors are increasingly using "cloud computing" as a marketing label for old technologies and offerings (cloud-washing), devaluing the term and the trend, and increasing the overall confusion in the IT industry. Although cloud computing indeed builds on "the shoulders of giants" — Gartner believes that its differentiation will amount to a transitional change in IT, not unlike the impact of the early Internet and the impact of the Web at a later time. The Internet connected individuals and organizations; the Web created the basis to inform individuals and organizations; cloud is evolving this phenomenon further to offer a platform for serving individuals and organizations. The tool that moved the industry from the early Internet to the Web was the browser (and its underlying HTTP/HTML standards); it was essentially the first introduction to the world of the principles of WOA. The transition to cloud too is enabled by a core architecture — SOA. Cloud services are designed for external and ubiquitous access — a fundamental objective of SOA principles of design. The use of registered interfaces and standard protocols ensures ubiquity of that access. SOA is not equivalent to Web services. Some of the cloud access will be via WS* protocols, but others will be via REST protocols (XML over basic HTTP) or event-driven SOA protocols. Cloud can be seen as cloud services, some exposed natively via SOA-style interfaces, others wrapped with user interfaces. Users invested in advanced SOA are best prepared to become users and providers of cloud-style services.
  • This presentation examines the nature of application infrastructure for cloud computing (PaaS) and looks at its current state, its architecture and road map, as well as at the impact the emerging PaaS technology and services will have on the enterprise IT practices and organization.
  • This presentation examines the nature of application infrastructure for cloud computing (PaaS) and looks at its current state, its architecture and road map, as well as at the impact the emerging PaaS technology and services will have on the enterprise IT practices and organization.
  • Software vendors across the spectrum of specialties are acutely aware of the emerging phenomenon of cloud computing. Although it is not entirely new (SaaS-style applications have been offered in some business contexts for years, and some level of infrastructure utility support — for storage and other purposes — has also been in use for a long time), the entry of new, visionary and powerful vendors into the space makes all the difference. Google and Amazon are two innovators that are leading the industry to the new cloud-based computing style. Cisco and WebEx have also contributed in this direction (WebEx Connect), though with lesser impact. More recently, IBM and Microsoft have begun multibillion-dollar investments in building data centers comparable to the likes of Google and Amazon to establish independent footprints for their own renditions of cloud computing. Both SAP and Oracle had pledged that the next generation of their business applications will be cloud-enabled and available (optionally) as SaaS. As in the data center — in the cloud, too — technology architecture is multilayered and complex. The basic hosting offerings (from IBM, Dell, Rackspace, OpSource and others) compete with the utility offerings from Amazon, Xdrive, MediaMax and others. The applications from Workday, NetSuite, RightNow and salesforce.com run over their own platforms, but many smaller and newer application ISVs build on the growing availability of development and run-time environments for the cloud, including Force.com, Rollbase, LongJump, Bungee Connect and others. With time, this variety of offerings will likely consolidate, but not in the immediate future, we expect a continuing emergence of new, innovative players and offerings.
  • Cloud is external to the user and elastic. It does not have to be external to the entire organization. When the cloud is managed inside an organization — it is private cloud. A private cloud is external opaque resource for the parts of the organization that own the applications that run there and it offers elasticity of resources — allocated dynamically to applications as needed, potentially scaling up and down swiftly and significantly on request. Private cloud is usually shared multiple domains of the organization's IT universe (divisions, departments, geographies). Public cloud is external to the entire organizations, managed and offered by third party providers. The resources are shared not only between applications and organization's domains, but also between many independent organizations — customers of the given cloud provider. Virtual private cloud is located outside of the organizations, but its resources are not shared with other organizations and the organization has full exclusive control of the characteristics and security arrangements. Community cloud is a cloud shared by the Government or a vertical industry: many agencies or enterprises, but still not open-to-all public model. Finally, a hybrid cloud is a combination of a private cloud and public cloud used in some kind of a relationship — for example the public cloud may be used for overflow of demand when the private cloud is overloaded or experiences downtime.
  • Cloud is external to the user and elastic. It does not have to be external to the entire organization. When the cloud is managed inside an organization — it is private cloud. A private cloud is external opaque resource for the parts of the organization that own the applications that run there and it offers elasticity of resources — allocated dynamically to applications as needed, potentially scaling up and down swiftly and significantly on request. Private cloud is usually shared multiple domains of the organization's IT universe (divisions, departments, geographies). Public cloud is external to the entire organizations, managed and offered by third party providers. The resources are shared not only between applications and organization's domains, but also between many independent organizations — customers of the given cloud provider. Virtual private cloud is located outside of the organizations, but its resources are not shared with other organizations and the organization has full exclusive control of the characteristics and security arrangements. Community cloud is a cloud shared by the Government or a vertical industry: many agencies or enterprises, but still not open-to-all public model. Finally, a hybrid cloud is a combination of a private cloud and public cloud used in some kind of a relationship — for example the public cloud may be used for overflow of demand when the private cloud is overloaded or experiences downtime.
  • The center of influence and standard setting in business IT has always been in the middle layer. The years of competition to dominate the platform technologies (Tuxedo versus Encina, .NET versus Java EE, Flash versus Silverlight) have resulted in the standardization of technologies, the creation of communities and ecosystems associated with one or the other platform technology, and the emergence of industry leaders. The application infrastructure (middle) layer established the programming model for applications above it (top layer) and, with maturity, allows a choice of system software and hardware (bottom layer) without affecting the applications. Customers committed to a certain application infrastructure architecture get to choose from a variety of competing applications and from a number of competing hardware options. This flexibility comes at the cost of lock-in to the chosen platform architecture (which, in some cases, translates into a lock-in to a vendor as well). Cloud-computing technologies are in their early stages of development. The cloud-style application infrastructure is only beginning to be offered, and some of the leading software vendors of today are not even in the market yet. However, we believe that the leadership of cloud computing will emerge with the vendors that invest and excel in providing the leading platform technology, architecture and standards for cloud services. Users are advised to watch the competition for the domination of the cloud application infrastructure (platform as a service [PaaS]) as the measure of the emerging maturity of the cloud-computing technology market.
  • In the traditional on-premises data center, support of business applications relies on application infrastructure technologies (also known as middleware) as the enabling technology for applications. This middle layer shields the applications from the technically challenging direct interactions from system software and hardware. Cloud architecture follows the same model. Business applications as a service (cloud application services) rely on underlying middleware that is cloud-enabled and serves as the foundation for the applications, shielding them from direct development of cloud-enabling functionality (such as the global-class scale, multitenancy, one-to-many self-service and others). In 2009 through 2011, the application infrastructure as a service market is fragmented and chaotic. Many vendors offer some components of the middleware suite as a service, and no one to date had stepped up to offer a complete, comprehensive and integrated middleware platform as a service (PaaS). The available options include such narrow offerings as the stand-alone application security or messaging middleware as a service (Veracode and StormMQ), as well as the broader application platforms as a service (Force.com and App Engine). Users today are called to act as system integrators, integrating the atomic middleware services, and are further burdened by the fact that each such service is offered from a different physical location. Use of these technologies today is not for everyone, and most are applied to relatively narrow-use scenarios.
  • In 2009 through 2011, the cloud application infrastructure services are offered by specialist vendors in a chaotic manner; many atomic functions are offered separately by different providers off different data centers. In an early cloud environment, this situation is tolerable, because the use of cloud middleware services is relatively narrow-focused and experimental. However, over time as organizations attempt to create more-significant cloud-based solutions, the pressure will be on providers to consolidate and offer targeted suites of cloud middleware services. By 2013, such suites will target prevailing use scenarios: new application development, integration of cloud/on-premises resources, information serving, governance and management of hybrid cloud/on-premises computing. The competing targeted suites will grow in their functional content, and by 2015, comprehensive single-source middleware service suites will emerge from the leading providers and their partner ecosystems. Users of cloud services will continue to choose between some narrow specialists, some targeted suites and several available compressive service platforms. Although special interest projects will often make unique choices — the mainstream IT users will tend toward providers with the greater communities of partners, add-on tools and business applications. Thus, the "war" for the PaaS domination is the war for application and tool-independent software vendors (ISVs).
  • The traditional mode of operation today is that the software of organization X runs on computers that belong to organization X and are managed by its IT department. There is one "tenant" (the organization itself), and it does not share its computing resources with others. The benefit of full control of integrity and the destiny of its strategic information resources comes with the burdens of responsibility for the same and of necessary (and costly) overcapacity to allow for spikes in demand. The cloud-computing model promises to address these and other costs. Four models are prevailing in implementing cloud-based business applications. The simplest form of cloud deployment is the one without multitenancy. Each organization gets its own resources, exclusively allocated just to it. There is no benefit of shared capacity, but many organizations use the same software solution, and they delegate the burdens of system management to the providers. The next model offers multitenancy based on OS virtualization. Each organization gets one or several virtual instances of the OS. The application and the application platform are unaware of the multitenancy. The standard application platforms and precloud applications run in this context without change. The currency of elasticity though is very coarse: It is an instance of an OS. The container level multitenancy allows multiple tenants to share the same computing resources at the application level and still preserve platform isolation. This model allows for the greatest elasticity at the cost of the most-proprietary environment. Finally, multitenancy can be built into the business application itself, but this requires the application designers to engage in system software issues, a risky proposition of dubious long-term value.
  • The traditional mode of operation today is that the software of organization X runs on computers that belong to organization X and are managed by its IT department. There is one "tenant" (the organization itself) and it does not share its computing resources with others. The benefit of full control of integrity and destiny of its strategic information resources come with the burdens of responsibility for the same and of necessary (and costly) overcapacity to allow for spikes in demand. The cloud computing model promises to address these and other costs. Four models are prevailing in implementing cloud-based business applications. The simplest form of cloud deployment is the one without multitenancy. Each organization gets its own resources, exclusively allocated just to it. There is no benefit of shared capacity, but many organizations do use the same software solution and they do delegate the burdens of systems management to the providers. The next up model offers multitenancy based on OS virtualization. Each organization gets one or several virtual instances of the OS. The application and the application platform are unaware of the multitenancy. The standard application platforms and pre-cloud applications run in this context without change. The currency of elasticity though is very coarse: it is an instance of an OS. The platform-level multitenancy allows multiple tenants to share the same computing resources at the application level and still preserve platform isolation. This model allows for greatest elasticity at the cost of the most proprietary environment. Finally, multitenancy can be built into the business application itself, but this requires the application designers to engage in systems software issues, a risky proposition of dubious long-term value.
  • The best practices for delivering or consuming cloud computing — and even the clarity of cloud computing's essential nature — are not yet settled. Vendors that have invested in the cloud each have a certain vision for what is the core value of cloud computing and how that value must be delivered to customers. However, there is no agreement between the vendors about these essentials. The recent confrontation between the CEOs of Oracle and salesforce.com during Oracle's OpenWorld event is just one example of such tension between the vendors. The vendors must choose between being a provider or an enabler of cloud computing, between multiple forms of multitenancy, between pricing models, between pure or hybrid cloud presence, and between buying markets (individuals, small and midsize businesses, large IT organizations, and ISVs). All have different requirements and expectations of cloud, and typically, one service model does not meet the expectations and needs of all types of buyers) Over time, leaders, standards and best practices will emerge, but today, users and prospects must understand that most of the technology architecture of cloud computing is in flux. Early virtualization offerings are likely to change in their aim to deliver more of the cloud value, proprietary multitenant platforms are likely focus on standardization, and many smaller players are likely to be acquired or change their business models to achieve sustainable business balance. Large players of today may redefine their architectures and visions over time, as their insight of the cloud opportunity and value increases. Clients can move to cloud with the tools and vendors that are available today, but they are well-advised to be prepared for change and some degree of discontinuity in the next three to five years, as the cloud computing industry grows and matures.
  • Salesforce.com is the undisputed early leader in APaaS/PaaS space. Its offering, Force.com, is largely based on the programmable foundation of Salesforce CRM— the company's CRM application enjoys more than 80,000 user organizations and more than 1 million named users. The experience of supporting and perfecting the platform for such a massive and increasingly complex installed base has helped salesforce.com to offer a highly scalable and well-thought-through application platform. In its first version, the platform is multitenant, and offers elastic scalability, fault tolerance, performance, Java-like 4GL programmability and rolling versioning. Salesforce competitors are small and experimental players, which is a challenge to its growth. To take APaaS/PaaS seriously, the mainstream users must see a vibrant market, including familiar software leaders (such as Oracle, IBM and Microsoft). Of those, Microsoft is committed to offering a competing offer. Oracle is moving in this direction, starting with private cloud-enabling technologies, but IBM so far has only made certain tentative and tactical moves toward articulating its vision for cloud computing. Oracle promises Fusion applications to be SaaS-enabled, which is likely going to push the company toward a serious APaaS/PaaS investment as well. Salesforce's own strategy is not without challenges. Its programming and runtime environments are proprietary, most of its users are small businesses, and it is struggling to break out of the market niche of a CRM vendor. An early leader, salesforce.com may be attractive as an acquisition target by a better-diversified major software industry player. Whatever the end-game, salesforce.com will likely help lead the industry out of the enterprise closet and into the open and agile era of cloud computing. Users can expect high productivity and scalability when using Force.com or Force-based business applications.
  • Microsoft offers a variety of off-premises products, from SharePoint online to Microsoft Dynamics CRM Online, Windows Live, Office Live and SQL Server Data services. By many accounts, Microsoft is under assault from Google and other Web innovators and is forced to adopt an off-premises computing model to meet those challenges. To address the changing computing environment and still preserve its core strength in client and server computing, Microsoft introduced the notion of software plus services. Under this model, the business-IT environment is a composition of both on-premises software and off-premises-sourced services.At year-end 2007, Microsoft announced a commitment to build four cloud-style massive data centers at a total cost of $2 billion — clearly a strategic investment on the same path. Microsoft has also announced and now delivered the early version of its Windows Azure cloud services environment. It is clear that Microsoft intends to be a major competitor in the APaaS/PaaS and cloud computing space, but it is also clear that delivering a tangible and consistent solution across its many product lines will take some time. In the process, users can assume that there will be changes to the product and business strategy of Microsoft, but they can also assume that Microsoft means it when it proclaims in its cloud advertising: "we are all in."
  • Oracle CEO, Larry Ellison, has spoken on multiple occasions to the effect that there is not a workable business model in SaaS software models. Nevertheless, at Oracle OpenWorld 2010, the company announced its entry into the cloud-computing space. In fact, its most significant new product — a middleware machine built on the template of its earlier Exadata offering — is called the "Exalogic Elastic Cloud" machine. Oracle promotes it as a cloud in a box, and some Oracle executives refer to it (as well as its sibling, Exadata) as a new type of mainframe — a consolidated, partly commodity and standards-based compute power that is designed to scale both up and out. However, the Exalogic machine in its present form is neither elastic nor cloud. It is simply an optimized rack assembly of compute power, without virtualization or any other support for elasticity and multitenancy. Improvements are promised for the future, but the name is used now — a confusing and, for many, misleading claim. Oracle promises that its forthcoming (1Q11) Fusion Applications will be cloud-enabled from the start. Oracle claims that these applications (including a CRM offering) will be offered from Oracle's data center in Houston, Texas, as well as by partners from other data centers around the world. This commitment puts Oracle in a direct competitive stance against salesaforce.com and will likely force the company to advance from rhetoric to the reality of cloud computing fast. At present, Oracle's stance on cloud is mostly to be an enabler to others that may want to offer a platform or an application as a service, and its first customers are IT organizations building private cloud arrangements. With time and with the growing maturity of Oracle Fusion Applications, this stance is likely to change. The salesforce company, challenged by Oracle in 2010, may become a close partner or more by 2012. Users and prospects must understand that Oracle is not yet ready to enable or deliver a cloud-computing experience, but it has thrown its hat in the ring, and much new technology and investment is bound to follow.
  • Google, a leading innovator in consumer applications and use patterns for Web and cloud computing, is a minimal player in enterprise systems. Its applications, such as Gmail, Docs and Search, are used in enterprises, but mostly are not considered enterprise applications. The applications do not provide multitenancy for tenant enterprises, because they are not offered to enterprises for exclusive use, isolated from other enterprises (tenants). That is not the business (or technology) model for Google. However, the Google computing environment offers multitenancy for tenant applications (multiple isolated applications elastically share common computing resources). The introduction of the Google App Engine started off in a similar consumer-developers-focused mode, offering only a Python programming environment (not an attractive option for enterprise projects). Recently, Google announced App Engine for Java, clearly indicating its interest in attracting enterprise developers. Despite the interest in the enterprise, Google is likely to remain a trailing competitor in enterprise application projects. It lacks in enterprise business experience, record or functional completeness. Third-party vendors, such as OrangeScape, may use App Engine as a core technology to offer an enterprise-multitenant platform. Nevertheless, as the new wave of technology architectures, driven by adoption of cloud computing and new metadata-driven programming models penetrates enterprises, some will see it as a strategic paradigm change and will embrace new, less enterprise-experienced vendors. Google might be a beneficiary of this trend. Until recently, Google has kept is internal system infrastructure secret. Naturally, none of it was offered for customers' use. Google does not offer any cloud system infrastructure services and, thus, is distinct from Amazon, which is often considered its cloud-computing competitor.
  • The opportunity to compete in the emerging software market of APaaS has inspired a number of vendors. Most are in the earliest stages of development (beta or early deployments). Force.com is the undisputed leader in the space, with the largest ISV and early customer community. However, even Force.com is less than two years old. Google is the largest overall vendor in the market, with its App Engine, although the product is still in beta and so far is limited in enterprise appeal without the support of a relational database model. Of the smaller vendors, dbFLEX has been in the market longest, but has achieved minimal name recognition. Rollbase and Bungee Labs (both still in beta) are differentiated by offering their platform technology as software products (not just as services). Rollbase also offers an open-source option where the buyer can fork the platform code and develop its own APaaS using the Bungee core. Rollbase, LongJump and Bungee use third-party data centers (OpSource, Rackspace or Amazon). Many vendors offer development tools aimed at mass market productivity. Some, in contrast, aim to attract advanced programmers and advanced projects. Some offer a computationally complete programming language. Not a single service offers the complete standard Java/Java EE programming model (although most are written in Java and use the Java EE Servlet model internally). Some use OS virtualization to instantiate tenants, and others have built multitenancy into the platform. The market is likely to run into an early conflict of objectives: the main focus on ease of use and productivity for modestly demanding projects vs. the essential requirement for high-end performance to support a growing community of tenants. Some of the early vendors may fail to meet both of these requirements.
  • Who: ISVs. The ISVs are key players in cloud computing, because the mainstream users of IT services are typically looking for solutions and reach for tools only if the solutions cannot be found. Moreover, the early adopters of cloud application services are typically SMB organizations. They are even more keen on looking for applications and would likely be resistant to developing new skills for application development. The ISVs providing cloud-based business applications are, thus, the key to much of early adoption of cloud computing.Why: The ISV business is much improved when applications are offered as a service, instead of the product (no multiple versions to support, no distribution costs, and no need to support multiple OSs, DBMSs and other factors). However, managing both the on-premises and the on-cloud options negates all of these benefits. ISVs also know that their SMB customers want the cloud option of the applications, and simply move there to meet the growing demand and because of competitive pressure.When: ISVs that target SMB organizations make the largest investment in cloud computing today (60% support cloud in part or in full) and will remain the leaders through 2015 (50% will be pure SaaS providers). Only 5% of ISVs targeting large systematic IT projects are pure SaaS today, increasing slowly to 12% by 2015.By 2015, more than half of SaaS applications will use a productized reusable CEAP or a PaaS.
  • Who: Application infrastructure (middleware) vendors. The middleware vendors will ultimately deliver the future standard (de facto or de jure) PaaS technologies. Their role in the market will be critical in establishing the mainstream use patterns of cloud computing. Why: The new application infrastructure innovators are investing strategically in cloud/PaaS to gain the early-entry advantage and to attract ISVs. The well-established middleware vendors aim to preserve the viability of the on-premises market where their strengths are. This tension will delay the all-out competition in the PaaS space, giving the early innovators further opportunities to establish their presence. When: All will recognize that attracting the ISVs is critical to the long-term success in this new market, and all will join the competition in the next three years. However, the early PaaS business will generate relatively low revenue compared with the traditional license revenue levels. That is, in part, because of the transition of capital to an operational cost basis, and in part because early competition is not about revenue but about influence, resulting in intentional minimal to negative margins. While today the vast majority of middleware revenue is not cloud-sourced (with the exception of some narrow areas, such as managed file transfer), by 2015, more than 25% of all new-business middleware revenue will be either direct PaaS revenue or revenue that includes partial support of PaaS.
  • Who: Enterprise IT projects (software development teams). When a packaged application or a cloud service is not available, the IT organization of the enterprise must develop the software itself. In the past, that development would have been entirely on-premises. In the future, increasing numbers will include cloud and, therefore, PaaS or CEAP, as the enabling technology.Why: SMB organizations, and smaller enterprise projects with limited budgets, increasingly look for advanced technology platforms to allow strong growth, high availability and scale to meet the requirements of the new global and growing business marketplace. Cloud platforms uniquely leverage the large community of small and large customers to offer state-of-the-art technology at SMB-level processes. Cloud platforms also reduce the capital costs, including reducing the demands on some internal IT resources. The larger IT organizations first choose the cloud platform for its exceptional time to the start of a project or because they envision a cloud-unique solution (involving social and other cloud-specific resources).When: Reduced-budget software projects (at SMBs, enterprise departments and elsewhere) are more likely to experiment or adopt a cloud-based technology platform. Of these projects, 40% will use cloud platform technologies (in part or in full) for their new projects. Large IT projects will continue to lag behind, but will be increasing their use of cloud platform technology as well, from 3% now, up to 20% by 2015.
  • Cloud-based services, especially the software (applications) as a service (SaaS) are offered as complete solutions and in most cases are packaged with user interfaces. However, even in the case of SaaS — a large share of access to cloud services is not via the user interfaces — but programmatic. Meaning that software built outside of a particular cloud service (on premises or in another cloud service) is using the cloud service as a component. This relationship of an application and a service is a service-oriented relationship and is best supported using the principles of service-oriented architecture (SOA). In fact, Gartner projects that the use of Cloud computing and cloud services will cement the use of SOA in mainstream enterprises in the next three years. The connectivity into the Cloud may be accomplished via Web Services using SOAP protocol, via REST using plain HTTP protocol or via event-posting protocols, typically over HTTP base as well. Contrary to some misguided positions, SOA is not equal to Web services and SOAP, but is an architectural model that is well-supported by multiple implementation protocols and patterns. The users best-equipped to use SOA principles in their IT projects are best prepared for cloud computing. The users invested in advanced SOA principles, including federates SOA deployments and event-driven SOA — will find cloud computing a natural extension to their established application infrastructure environment. Application developers, considering offering their application as cloud services (immediately or in a potential future) must follow SOA principles to assure cloud-style access to application.
  • The traditional enterprise is monolithic. There is no distinction between the front and back ends of the applications — everything runs on one platform. Applications are isolated to ensure the homogeneity and the optimization and relative simplicity of management that comes with it. A service-oriented enterprise is heterogeneous. A user-facing application logic calls on the back-end resources of multiple applications, and most user experiences involve integration, composition and, most recently, the mashup of information and transaction resources. SOA increases agility, but it also increases the burdens of the enterprise in managing the multiple platforms, architectures, products and protocols. Integration infrastructure is essential, and every enterprise struggles to build their own, although all are addressing essentially the same problem. Cloud computing emerges as the product of the maturing Web, SOA and integration. Applications in the cloud are managed on behalf of enterprises, relieving some of their system management burdens. Although some enterprises do not yet trust the cloud and third parties to protect their data and services, this is likely only a matter of time. As cloud content and credibility grows, enterprise computing will become increasingly dependent on it. Cloud, in turn, will introduce standards for systems management, assembly, security and quality of service that will lead the software industry overall to the next stage of its industrialization: reusable predictable parts, protocols and interfaces; predictable delivery schedules and predictable quality. The exchanges in the cloud will likely emerge to ensure integration as a service qualities. The B2B Gateway technologies and ESB suites are the likely precursors for these exchanges.
  • Df10200 gartner

    1. 1. Yefim Natis<br />Defining The Cloud Platform<br />
    2. 2. Cloud Computing: Evolution Inside a Revolution<br />SOA<br />serving<br />Browser<br />informing<br />connecting<br />Cloud<br />Web<br />Internet<br />1980<br />1990<br />2000<br />2010<br />2020<br />
    3. 3. Key Issues<br />What is PaaS, its architecture and its road map?<br />How will the vendors' strategies influence the PaaS road map?<br />PaaS road map: Who, why and when?<br />
    4. 4. Key Issues<br />What is PaaS, its architecture and its road map?<br />How will the vendors' strategies influence the PaaS road map?<br />PaaS road map: Who, why and when?<br />
    5. 5. Elastic, shared Internet resources<br />Fixed, dedicated resources<br />Cloud<br />Hosting<br />Hardwaremanagedby others<br />Off-Premises<br />The Cloud Architecture Is a Three-Layer Cake<br />V-Cloud<br />Application Services (SaaS*)<br />Application Infrastructure Services (PaaS*)<br />System Infrastructure Services (IaaS*)<br />* According to NIST (National Institute of Standards and Technology), http://csrc.nist.gov/groups/SNS/cloud-computing/index.html<br />
    6. 6. Private vs. Public: Understanding the Trade-Offs<br />Private Cloud<br /><ul><li>Designated enterprise data center (or segment) managed centrally
    7. 7. Data center resources shared by all divisions, protected by enterprise central controls
    8. 8. Divisions of enterprise act as independent tenants
    9. 9. Some elasticity of resources; good resource utilization; reduced cost of business</li></ul>Enterprise IT<br /><ul><li>Each enterprise division manages its own data center (or a subdivision)
    10. 10. Exclusive local control of resources
    11. 11. Internally borne costs and burdens of management
    12. 12. High-cost overcapacity, low resource utilization</li></ul>Virtual Private Cloud<br /><ul><li>Third-party data center providers (public cloud characteristic)
    13. 13. Data center sharing is restricted to only the divisions of this enterprise (private cloud characteristic)
    14. 14. Divisions of enterprise act as independent tenants (private cloud characteristic)
    15. 15. Some elasticity; good resource utilization; low cost of business</li></ul>Community Cloud<br /><ul><li>Consortium or a government scope data center (larger than private, but smaller than public)
    16. 16. Members of the consortium or government agencies act as independent tenants
    17. 17. Data center resources are shared by all members; consortium provides security, privacy and capacity
    18. 18. Good elasticity of resources; high resource utilization; reduced cost of business</li></ul>Public Cloud<br /><ul><li>Third-party data center providers
    19. 19. Computing resources shared by independent enterprises (tenants), protected by third parties in cloud
    20. 20. Maximum elasticity; maximum resource utilization; low cost of business</li></ul>Public Cloud<br />Private Cloud<br />Virtual Private Cloud<br />Community Cloud<br />No Cloud<br />Enterprise 1<br />Enterprise 2<br />
    21. 21. Private vs. Public: Understanding the Trade-Offs<br />Cost-Efficiency<br />Autonomy<br />Private Cloud<br /><ul><li>Designated enterprise data center (or segment) managed centrally
    22. 22. Data center resources shared by all divisions, protected by enterprise central controls
    23. 23. Divisions of enterprise act as independent tenants
    24. 24. Some elasticity of resources; good resource utilization; reduced cost of business</li></ul>Enterprise IT<br /><ul><li>Each enterprise division manages its own data center (or a subdivision)
    25. 25. Exclusive local control of resources
    26. 26. Internally borne costs and burdens of management
    27. 27. High-cost overcapacity, low resource utilization</li></ul>Virtual Private Cloud<br /><ul><li>Third-party data center providers (public cloud characteristic)
    28. 28. Data center sharing is restricted to only the divisions of this enterprise (private cloud characteristic)
    29. 29. Divisions of enterprise act as independent tenants (private cloud characteristic)
    30. 30. Some elasticity; good resource utilization; low cost of business</li></ul>Community Cloud<br /><ul><li>Consortium or a government scope data center (larger than private, but smaller than public)
    31. 31. Members of the consortium or government agencies act as independent tenants
    32. 32. Data center resources are shared by all members; consortium provides security, privacy and capacity
    33. 33. Good elasticity of resources; high resource utilization; reduced cost of business</li></ul>Public Cloud<br /><ul><li>Third-party data center providers
    34. 34. Computing resources shared by independent enterprises (tenants), protected by third parties in cloud
    35. 35. Maximum elasticity; maximum resource utilization; low cost of business</li></ul>Public Cloud<br />Private Cloud<br />Virtual Private Cloud<br />Community Cloud<br />No Cloud<br />Enterprise 1<br />Enterprise 2<br />
    36. 36. <ul><li>By 2013, all leading software vendors will aggressively compete for leadership in the cloud platform market.
    37. 37. Through 2013, the cloud platform leadership will be with PaaS providers that have the largest ISV ecosystems
    38. 38. By 2015, cloud platform experience will be a listed or demanded skill in most hiring decisions by IT software projects.</li></ul>Platform as a Service: The Strategic Center of Cloud-Computing Architecture<br />Applications<br />SaaS<br />Application Infrastructure<br />PaaS<br />Programming models, languages, standards, interoperability, architectures, developer skills, partner ecosystems, user communities, market influence<br />IaaS<br />System Infrastructure<br />
    39. 39. Cloud Computing:Valuable Services, but Chaos Reigns<br />App Engine<br />Integrated Application Servers<br />Data Cloud<br />Click-to-secure Cloud<br />Atmosphere<br />Cloud Services<br />Application Security<br />Data Integration<br />Collaborative network<br />Managed File Transfer (MFT)<br />Apps Marketplace<br />Application and B2B Integration<br />DBMS File Systems<br />XAP-aaS<br />App. Marketplaces (catalogs)<br />Cloud IQ<br />Covicint Portal<br />eXtreme TP<br />App. Life Cycle Management (ALM)<br />Portals UXP<br />RDS, Simple DB<br />Messaging Middleware<br />BPM Technology<br />Cloud<br />Interstage<br />Agile planner<br />Based on "Application Infrastructure for Cloud Computing: a Growing Market, 2010" G00175138 <br />
    40. 40. PaaS 2015: Comprehensive Application Infrastructure Platform as a Service<br />Application Infrastructure Services<br />Targeted Application Infrastructure Platforms as a Service<br />Application Security<br />Integrated Application Servers<br />Application Platform Services<br />Application Platform<br />as a service<br />(aPaaS) <br />DBMS, File Systems<br />Application and B2B Integration<br />Information Platform Services<br />Data Integration<br />Managed File Transfer<br />Comprehensive Application Infrastructure Platform as a service<br />(PaaS)<br />Portal/User Experience<br />Business Process Management<br />Integration Platform Services<br />Management/Governance<br />Integration Platform<br />as a service<br />(iPaaS) <br />Extreme TP Application Server<br />Messaging Middleware<br />Governance Platform Services<br />Application Life Cycle Management<br />2015<br />App. Marketplace (Catalog)<br />2013<br />2011<br />
    41. 41. Gartner Reference Architecture for Multitenancy <br />3<br />Shared OS<br />1<br />SharedNothing<br />2<br />Shared Hardware<br />6<br />Shared Everything<br />7<br />Custom Multitenancy<br />4<br />Shared Database<br />5 <br />Shared Container<br />Tenant<br />Tenant<br />Tenant<br />Tenant<br />Tenant<br />Tenant<br />Tenant<br />Tenant<br />Tenant<br />Tenant<br />Tenant<br />Tenant<br />Tenant<br />Tenant<br />App.<br />App.<br />App.<br />App.<br />App.<br />App.<br />Application Logic<br />Cloud-Enabled Application Logic<br />Application Logic<br />App.<br />App.<br />ApplicationLogic<br />AP<br />AP<br />Cloud-Enabled Application Platform <br />AP<br />AP<br />AP<br />AP<br />Cloud-Enabled Application Platform <br />Application Platform<br />AP<br />AP<br />ApplicationPlatform<br />DP<br />DP<br />DP<br />DP<br />DP<br />DP<br />DP<br />DP<br />Cloud-Enabled Data Platform<br />Data Platform<br />Cloud-Enabled Data Platform<br />Data Platform<br />Sys. Inf.<br />Sys. Inf.<br />System Infrastructure<br />System Infrastructure<br />System Infrastructure<br />System Infrastructure<br />System Infrastructure<br />System Infrastructure<br />SystemInfrastructure<br />Managed Virtual Machines<br />Managed<br />OS Processes<br />Multitenancy<br />From “Gartner Reference Architecture for Multi-tenancy“ G00205983<br />
    42. 42. The Cloud Architecture War: Evolution or Revolution?<br />Users<br />Users<br />Tenant <br />Tenant <br />Tenant <br />Tenant <br />Application <br />Application <br />Application <br />Application <br />Instance<br />Instance<br />Instance<br />Instance<br />Application <br />Application <br />Application <br />Application <br />Application <br />Application <br />Application <br />Application <br />Server<br />Server<br />Server<br />Server<br />Server<br />Server<br />Server<br />Server<br />Business <br />Business <br />Business <br />Business <br />Business <br />Business <br />Business <br />Business <br />Data<br />Data<br />Data<br />Data<br />Data<br />Data<br />Data<br />Data<br />Elastic Management Infrastructure<br />Elastic Management Infrastructure<br />Elastic Management Infrastructure<br />Elastic Management Infrastructure<br />Operating System<br />Operating System<br />Operating System<br />Operating System<br />Virtualization Infrastructure<br />Virtualization Infrastructure<br />Virtualization Infrastructure<br />Virtualization Infrastructure<br />Hardware Grid<br />Hardware Grid<br />Hardware Grid<br />Hardware Grid<br />Shared<br />-<br />hardware <br />Shared<br />-<br />hardware <br />Shared<br />-<br />hardware <br />Shared<br />-<br />hardware <br />Multi<br />-<br />tenancy<br />Multi<br />-<br />tenancy<br />Multi<br />-<br />tenancy<br />Multi<br />-<br />tenancy<br />Users<br />Users<br />Tenant <br />Tenant <br />Tenant <br />Tenant <br />Application <br />Application <br />Application <br />Application <br />Instance<br />Instance<br />Instance<br />Instance<br />Application<br />Application<br />Multi<br />-<br />tenancy <br />Multi<br />tenancy <br />Multi<br />-<br />tenancy <br />Multi<br />tenancy <br />Control<br />Control<br />Control<br />Control<br />SaaS/Cloud<br />-<br />enabled <br />SaaS/Cloud<br />-<br />enabled <br />SaaS/Cloud<br />-<br />enabled <br />SaaS/Cloud<br />-<br />enabled <br />Application Server<br />Application Server<br />Application Server<br />Application Server<br />Business <br />Business <br />Business <br />Business <br />Data<br />Data<br />Data<br />Data<br />Operating System<br />Operating System<br />Operating System<br />Operating System<br />Hardware Grid<br />Hardware Grid<br />Hardware Grid<br />Hardware Grid<br />Shared<br />-<br />everything <br />Shared<br />-<br />everything <br />Shared<br />-<br />everything <br />Shared<br />-<br />everything <br />Multi<br />-<br />tenancy<br />Multi<br />-<br />tenancy<br />Multi<br />Multi<br />
    43. 43. Key Issues<br />What is PaaS, its architecture and its road map?<br />How will the vendors' strategies influence the PaaS road map?<br />PaaS road map: Who, why and when?<br />
    44. 44. The Battle of Visions: Vendor Strategies Under Construction<br />Provider<br />Enabler<br />Private<br />Public<br />Shared-Hardware<br />Shared-Everything<br />Java/C#<br />Cloud App. Dev.<br />Pay-per-Use<br />Fixed-Price<br />Pure Cloud<br />Hybrid<br />IT Buyers<br />ISV Buyers<br />
    45. 45. Salesforce.com: The Enterprise Cloud Platform Market Is Its to Lose<br /><ul><li>Provider
    46. 46. Public cloud
    47. 47. Shared-everything
    48. 48. Dedicated cloud programming
    49. 49. Fixed-price</li></ul>Application Services<br />Application Infrastructure Services<br />System Infrastructure Services<br />?<br />(future)<br />
    50. 50. Application Services<br />Application Infrastructure Services<br />System Infrastructure Services<br />Microsoft: Taking the Battle to the Clouds<br /><ul><li>Enabler and provider
    51. 51. Public and private cloud
    52. 52. Shared-hardware
    53. 53. Traditional programming
    54. 54. Fixed- and use-price</li></ul>VM Role<br />platform<br />
    55. 55. Oracle: Cloud Lives in a Box<br /><ul><li>Enabler
    56. 56. Private and public cloud
    57. 57. Shared-hardware
    58. 58. Traditional programming
    59. 59. Fixed-price</li></ul>CRM On-demand<br />Oracle Fusion Applications (future)<br />Application Services<br />Application Infrastructure Services<br />System Infrastructure Services<br />Enterprise Edition 7 (future)<br />12c<br />(future)<br />
    60. 60. Application Services<br />Application Infrastructure Services<br />System Infrastructure Services<br />Google: Consumer Cloud Discovers the Enterprise<br /><ul><li>Provider
    61. 61. Public cloud
    62. 62. Shared-database and shared-everything
    63. 63. Traditional programming and cloud programming
    64. 64. Use-price</li></ul>Calendar<br />Sites<br />?<br />
    65. 65. Who's Who in Application Platforms for Cloud Computing <br />Enterprise Generalists<br />Cloud Specialists<br />Adobe<br />Appistry<br />Aptana<br />Cordys<br />GigaSpaces<br />IBM<br />Intuit<br />Kovair<br />maatG <br />Magic Software<br />Micro Focus<br />Microsoft<br />Oracle<br />Progress Software<br />Rackspace<br />Red Hat<br />VMware SpringSource<br />Tibco Silver<br />WaveMaker<br />Apprenda<br />Bungee Labs <br />Caspio<br />Corent<br />Engine Yard<br />Express Dynamics WorkXpress<br />ForeSoft dbFLEX<br />Google App Engine<br />Heroku<br />OrangeScape<br />Qrimp<br />Relationals (LongJump)<br />Rollbase<br />salesforce.com Force.com<br />SiteMasher<br />Stax<br />TrackVia<br />VMware/salesforce.com VMforce<br />Zoho<br />
    66. 66. Key Issues<br />What is PaaS, its architecture and its road map?<br />How will the vendors' strategies influence the PaaS road map?<br />PaaS road map: Who, why and when?<br />
    67. 67. PaaS Road Map: Who, Why and When?1. Independent Software Vendors<br /><ul><li> To reduce the cost of operations, version control and support
    68. 68. To help SMBs reduce their dependence on internal IT
    69. 69. To offer state-of-the art technology at SMB prices
    70. 70. To help enterprises off-load commodity computing</li></ul>50%<br />Share of ISVs that are pure SaaS providers<br />50%<br />50%<br />30%<br />20%<br />25%<br />15%<br />PaaS CEAP<br />12%<br />7%<br />5%<br />Custom Platform<br />2015<br />2013<br />2011<br />ISVs targeting large enterprise IT <br />ISVs targeting SMBs and small/departmental IT projects<br />
    71. 71. PaaS Road Map: Who, Why and When?2. Application Infrastructure Vendors<br /><ul><li>To retain its customer base as its business-IT practices change toward greater expediency, agility and productivity
    72. 72. To enter a formerly saturated middleware market with a competitively radical change
    73. 73. To attract an ecosystem of partner ISVs and Sis
    74. 74. To reduce operation costs, costs of version control, distribution and support</li></ul>Share of AIM market new business revenue from cloud-related projects<br />(AIM = application infrastructure and middleware)<br />50%<br />25%<br />12%<br />5%<br />2011<br />2015<br />2013<br />
    75. 75. PaaS Road Map: Who, Why and When?3. Enterprise IT Projects<br /><ul><li> To get access to state-of-the art platform technology at SMB prices
    76. 76. To reduce or eliminate dependence on internal IT
    77. 77. To reduce capital costs and gain productivity
    78. 78. To salvage a late project
    79. 79. To build unique cloud applications</li></ul>Share of new custom application projects that use, in part or in whole, a PaaS for runtime deployment<br />40%<br />50%<br />20%<br />20%<br />Pure cloud<br />10%<br />5%<br />3%<br />2011<br />2015<br />2013<br />Hybrid cloud and on-premises<br />SMBs, departmental projects, small projects<br />Large enterprise IT projects<br />
    80. 80. Key Issues<br />What is PaaS, its architecture and its road map?<br />How will the vendors' strategies influence the PaaS road map?<br />PaaS road map: Who, why and when?<br />In conclusion<br />
    81. 81. Users<br />Cloud Will Cement SOA in Mainstream IT<br />Applications<br />Application Services<br />SOA Services (WS*, REST, Events)<br />Application Infrastructure Services<br />System Infrastructure Services<br />
    82. 82. Cloud Applications<br />Cloud Services<br />SOA <br />User-Serving Front End<br />Cloud-Aware Enterprise: Three Platforms, Integrated by the Fourth<br />Enterprise Core Back End<br />
    83. 83. Your Cloud Platform Action Plan<br />Monday Morning:<br />Become familiar with some of the cloud computing an PaaS options. Try a small, low-risk APaaS-based application to get a sense of the new environment.<br />The Next 12 Months:<br />Look for opportunities of low risk to make some real contributionto your business using cloud-computing services and APaaS.<br />Stay informed about new developments. Large software vendors will likely enter the field in some notable ways and will have to position their cloud strategies next to their traditional on-premises software strategy.<br />Continue to invest in SOA-style applications and infrastructure to be ready for cloud computing.<br />On the Radar Screen:<br />Prepare to use PaaS as a complement to traditional on-premises middleware tools — for notable business application projects.<br />Continue to strengthen your ability to integrate and compose applications built of heterogeneous and multisourced components across cloud and on-premises applications, information and tools.<br />
    84. 84. Related Gartner Research<br /><ul><li>Gartner Reference Architecture for MultitenancyYefim Natis and Erip Knipp (G00205983)
    85. 85. Gartner Reference Architecture for Cloud-Enabled Application PlatformsYefim Natis, Massimo Pezzini, Eric Knipp (G00201437)
    86. 86. VMware and Salesforce.com: The Beginning of a Beautiful Friendship?Yefim Natis and others (G00200632)
    87. 87. Application Infrastructure for Cloud Computing: A Growing Market, 2010Yefim Natis and others (G00175138)
    88. 88. Microsoft AppFabric: A Platform for the Cloud Era Is Under ConstructionYefim Natis, Massimo Pezzini, David Cearley, Eric Knipp (G00173880)
    89. 89. Creating Cloud Solutions: A Decision FrameworkEric Knipp, Yefim Natis and others (G00171623)</li></ul>For more information, stop by Gartner Solution Central or e-mail us at solutioncentral@gartner.com. <br />
    90. 90. Yefim Natis<br />Defining The Cloud Platform<br />

    ×