Live News, Features and Article Updates Covering Oracle Technologies

Oracle on Ulitzer

Subscribe to Oracle on Ulitzer: eMailAlertsEmail Alerts newslettersWeekly Newsletters
Get Oracle on Ulitzer: homepageHomepage mobileMobile rssRSS facebookFacebook twitterTwitter linkedinLinkedIn


Oracle Authors: Elizabeth White, Yakov Fain, Liz McMillan, Pat Romanski, Carmen Gonzalez

Related Topics: Cloud Computing, Enterprise Architecture, Oracle Journal, SOA & WOA Magazine, Microservices Journal, Cloud Development Tools

Article

Beware of the Accidental Cloud Architecture

An exclusive Q&A with Bruce Tierney, a Director for Oracle SOA Suite at Oracle

Prior to service-oriented architecture (SOA), many companies suffered from the infamous "accidental architecture" which evolved as a result of accumulating a large number of point-to-point application integration connections. Unless IT takes a strong role ensuring Software as a Service (SaaS) purchases by lines of business (LOB) integrate easily with the existing infrastructure, history is likely to repeat itself by reversing the benefits of SOA with a new form of chaos called the "accidental cloud architecture"

In this Q&A, Oracle's Bruce Tierney, a director of Oracle SOA Suite, will outline a number of best practices businesses can adopt to avoid creating the accidental cloud architecture. Bruce will also explain how organizations can best leverage SOA and BPM technologies to create an optimal hybrid on-premise/private/public cloud infrastructure.

Q: What is the "Accidental Cloud Architecture" and why should we beware?

Bruce Tierney: The painful memory of large-scale point-to-point application integration complexity is starting to fade as companies transition to a service-based integration (SOA). The "accidental architecture" which evolved as a result of accumulating these point-to-point connections occurred well before SOA and cloud. Then, SOA came to the rescue, allowing IT departments to gradually shift to an improved architecture with applications connecting to an Enterprise Service Bus (ESB) instead of directly to each other. However, SaaS has dramatically changed the dynamics between IT and LOB for software purchases providing opportunities for improvement or a path to chaos, depending on how strong a role IT plays.

Just as organizations are closing in on the benefits of service-based integration, along comes the cloud-enabled ability for lines-of-business to outsource application installation and more to SaaS vendors, reducing the dependency they had on IT but not always for the good of the organization. LOBs have the opportunity to move faster with new acquisition purchases by simply bypassing the IT Center of Excellence (CoE) best practice governance. If IT is not actively engaged with cloud software acquisitions, the result is the evolution of an architecture that was not planned but grew ad-hoc by incrementally adding one new connection to a cloud vendor after another into a new "accidental cloud architecture."

Q: Can you provide specific scenarios that illustrate the kinds of business problems this "accident" can cause?

Tierney: When the role of the IT organization is minimized or bypassed during the acquisition of SaaS, a number of potentially far-reaching problems can occur. First, what was initially not an IT acquisition decision will become IT's problem if challenges occur when the mostly siloed SaaS application needs to be integrated with the rest of the on-premise infrastructure. Secondly, ignoring the Center of Excellence best practices such as governance, service policies and more becomes a serious risk to the reliability and performance of the infrastructure. Further, multiple development environments, business rules, services buses, etc. become dispersed across multiple clouds, which eliminates the possibility of economies of scale when consolidating to one integration infrastructure. And finally, and perhaps most obviously, security could become a major challenge. The service provider uses security policies that may not be the same as your infrastructure and if an issue occurs, your customer will blame your company even if the cause was the cloud vendor your company outsourced to, not your company. Service Level Agreements (SLAs) between your cloud vendor and your company is a safeguard against this risk, but keep in mind that SLAs are agreements, not guarantees.

Q: How do organizations avoid creating an accidental cloud architecture?

Tierney: The good news is that it's still early on in the transition from on-premise to hybrid on-premise/cloud. If IT takes proactive steps soon, the accidental cloud architecture can be avoided. There are essentially four steps and best practices that IT managers should take to ensure they are not the victims of an accidental cloud architecture:

  • Executive buy-in: The accidental cloud architecture needs to be understood at the executive level. That means executive buy-in on the concept that IT is an approver of SaaS acquisitions, including being involved in or signing off on SLAs, to ensure the integration approach mandated by the SaaS vendor works well with Center of Excellence best practices. It also means making CXOs aware that there is a process in place for architecture decisions, and SaaS acquisitions by LOBs are to be included in this process. IT must take the initiative to get executive buy-in in this evolving hybrid model or risk being marginalized initially and blamed later.
  • Strong center of excellence (CoE): It's important that the CoE (or CoEs, as some organizations employ multiple global teams) is viable, effective, and recognized within the organization. Otherwise, the expansion into a more distributed hybrid cloud infrastructure will result in management chaos, reduced reliability, increased complexity and an inability to evaluate ROI.
  • Technology: It is very beneficial to have a SOA in place before an organization moves to the cloud. Though companies may not need to transition to 100 percent SOA, the aggressive introduction of the cloud brings urgency to the SOA model. Another important point about the SOA horse before the cloud cart: as organizations form partnerships with cloud vendors, it's important that they have a low barrier to exit and the flexibility to pull back or to move to another cloud vendor. This flexibility can be achieved with loose coupling between an organization and its service providers. Likewise, as business requirements change over time it's likely there will be the need for more choice in deployment style. For example, if new regulatory requirements enforce stronger penalties for leaking customer information, there may be a desire to migrate the deployment from cloud back to on-premise or private cloud.
  • Business process management (BPM): BPM is the next wave following SOA, and can be a great benefit to the emergence and development of the hybrid cloud. Once an organization gains executive buy-in for its CoE, it can create a business process around cloud-service requests, approval, and confirmation. This ensures that IT is a key player in that decision-making process.

Q: Will cloud be the ultimate end-state for future deployments?

Tierney: Although companies are steadily deploying more SaaS applications, many companies may want to avoid static end-state for their deployments. Some new business requirements will demand deployment flexibility so portions of the infrastructure hosted on-premise may move to a private cloud if rapid scalability becomes critical. Likewise, other portions of the infrastructure may move from public cloud to on-premise for tighter control over all aspects of the platform and infrastructure, including the operating system. So the desirable end-state for many companies will be deployment flexibility and choice.

Q: What roles do governance and API management play in managing a hybrid cloud?

Tierney: In the context of a hybrid deployment architecture, governance is an even more critical concept than in the on-premise world - and when it falls by the wayside, it can create another major business problem in addition to the various scenarios previously mentioned. The hybrid cloud represents a potentially significant expansion in infrastructure heterogeneity by introducing new protocols, standards and proprietary formats across all of an organization's new cloud partners that are now an integral part of its network.

API management, which allows organizations to remove barriers for externally leveraging service assets, and the hybrid cloud have a strong correlation, too. The distributed nature of the infrastructure due to hybrid deployments drives organizations to take API management seriously. Companies wishing to provide a unified API across disparate deployments need an API management solution to provide a cohesive user experience. Based on these realities, both governance and API management become stronger requirements for hybrid deployment architectures than in the purely on-premise world.

Q: Can you provide an example of how a customer would benefit from a hybrid cloud architecture optimized for SOA and BPM?

Tierney: Imagine a scenario where SaaS CRM was used exclusively by the marketing department, sales uses a legacy CRM on-premise application, and a company with yet another SaaS CRM product is acquired and merged into the now expanded enterprise. A Customer Service Representative (CSR) takes a call from a customer who exists in all three systems. In the ideal scenario, using a combination of Master Data Management, SOA, and BPM, the CSR gets a complete view of the customer's information and can update the respective applications and data stores. If the application embedded integration and process tools from each application were cobbled together to span all three of these, it's easy to visualize the complexities building a process leveraging all three toolkits, as well as the added costs of maintaining expertise in all three integration and process toolkits. The optimized solution for BPM would be a shared foundation with a SOA-based integration platform, to avoid redundant styles of connectivity (adapters, Web services, B2B, etc.) for BPM versus SOA. Likewise, the SOA-based integration platform should eliminate the point-to-point connectivity common in these scenarios with a single enterprise service bus. As more cloud vendors are added to the mix over time, each mandating their chosen standards (and non-standards), so too is the increase in the requirement for a platform optimized to handle any-to-any integration formats.

Q: What are some of the inherent challenges in adopting a hybrid cloud model?

Tierney: What was once managed by people with the same company badge as you is now being managed according to SLAs by people you may never contact in person or by any electronic means. So, what's required is a new skillset for developing and negotiating these SLAs, to make sure they get the reliability and quality of service on par or better than before the hybrid evolution. Likewise, the role of internal IT shifts as less operational support is needed with a transition to more SaaS applications, yet new architectural skills are required to blend the increasingly diverse composite cloud partner network into a unified experience for your end users, be they employees or customers.

Q: What kind of SOA and BPM solutions does Oracle offer, and how does Oracle differentiate itself in the integration/SOA market?

Tierney: Oracle SOA Suite differentiates from other integration platforms in the market in several ways. Firstly, rather than adding multiple niche integration toolsets for each new integration problem, Oracle SOA Suite has the breadth and depth to support the increasingly diverse integration requirements as each new cloud vendor is added to the network. Secondly, Oracle SOA Suite provides a single, unified experience spanning all application deployment scenarios and all of the integration components used during the transaction of an integration process. And finally, Oracle BPM Suite layers directly on top of Oracle SOA Suite so there is no duplication of integration components across BPM and SOA.


Bruce Tierney is a Director for Oracle SOA Suite at Oracle. He is a co-author of Cloud Integration - A Comprehensive Solution, author of Rethink SOA: A Recipe for Business Transformation and has delivered SOA and business integration training globally to help companies optimize and simplify business integration. Bruce's specialties include Service Oriented Architecture (SOA), Enterprise Service Buses (ESB), Business Process Management (BPM), and Business-to-Business Integration (B2B). Follow Bruce @tierneybruce

To learn more about SOA keynotes and sessions at Oracle OpenWorld 2013, visit: https://oracleus.activeevents.com/2013/connect/focusOnDoc.do?focusID=23446

More Stories By Elizabeth White

News Desk compiles and publishes breaking news stories, press releases and latest news articles as they happen.

Comments (0)

Share your thoughts on this story.

Add your comment
You must be signed in to add a comment. Sign-in | Register

In accordance with our Comment Policy, we encourage comments that are on topic, relevant and to-the-point. We will remove comments that include profanity, personal attacks, racial slurs, threats of violence, or other inappropriate material that violates our Terms and Conditions, and will block users who make repeated violations. We ask all readers to expect diversity of opinion and to treat one another with dignity and respect.