Q&A: BI and Cloud Computing

Why the impact of cloud BI is unavoidable.

• By Linda L. Briggs
• 04/27/2011

With all the talk about business intelligence in the cloud and real-time cloud integration, when should companies consider applications in the cloud? In this interview, Darren Cunningham, VP of cloud data integration marketing at Informatica, discusses some of the basics and impacts of cloud computing. Cunningham is responsible for leading product, customer and programs marketing for Informatica Cloud, the company’s on-demand data-integration-as-a-service solution.

The impact of cloud computing is broad and deep, Cunningham says -- the sort of technology that companies will need to either embrace or be run over by. Before joining Informatica, Cunningham was VP of marketing at LucidEra, an early entrant in the software-as-a-service (SaaS) business intelligence market. Prior to LucidEra, he led analytics product management and AppExchange data management partnerships at Salesforce.com. Cunningham also spent over seven years at Business Objects, where he managed product marketing for many of the company’s primary business intelligence, data integration, operational reporting, and performance management solutions.

BI This Week: What impact is cloud computing having on traditional IT roles and responsibilities?

Darren Cunningham: I spoke with a CIO recently who put it this way: “Either we get on the cloud train or we get run over by it.” I think that’s a good summary.

Some of the key benefits of cloud computing include rapid implementation, ease of use, and subscription pricing. Not surprisingly, in many organizations, the potential of computing in the cloud has led to empowered lines of business that don’t want to wait for permission from centralized IT to select, implement, and administer cloud applications and platforms. The IT command-and-control model of the past will no longer be acceptable in the new reality of a hybrid cloud. IT organizations must work closer than ever with the business to ensure they’re not only adding value but helping to drive business innovation and competitive differentiation.

What are some of the barriers to adoption of cloud-based applications and platforms?

Data privacy and security are always cited as the primary barriers to cloud adoption -- whether it’s software-, platform-, or infrastructure-as-a-service. This is an obvious concern in many ways as organizations consider moving critical business data off premises. However, as technologies, vendors, buyers, and overall implementations have matured, having the right data integration strategy has emerged as the key driver of cloud success. In reality, many organizations are now struggling with disconnected “SaaS marts” and “SaaS sprawl.” The result is fragmented enterprise data, limited ROI, and multiple versions of the truth.

What is the role of data integration as it relates to cloud computing?

The primary cloud data integration use cases we see are data migration (from legacy systems), data synchronization between applications (“cloud to ground” and “cloud to cloud”), data quality, and data replication (often to a data warehouse for business intelligence purposes). On the infrastructure side, there is also strong demand for cloud data archival and storage. The good news is that there is now broad recognition that “No Software” does not mean “No Data Integration,” but it may mean new approaches are necessary to improve collaboration between IT and business.

Can companies move incrementally to the cloud without making a full-on commitment?

Absolutely. It’s exciting to see many companies (and even the federal government) adopt “cloud-first” strategies, but the reality is that we now live in a hybrid IT world. The cloud represents a great opportunity to focus on your core competencies and dramatically reduce IT infrastructure costs, but don’t lose sight of what’s most critical -- your data. Companies should start by doing an audit of what SaaS applications they already own (this may shock them) and what data integration tools are being used (if any). Consider creating an enterprise architecture function responsible for determining what will move to the cloud, how, and when, and seek ways to empower business groups with self-service in a way that won’t result in downstream data disasters.

What about service-level agreements? How should they be enforced with cloud vendors?

Less-mature cloud vendors are not always proactive at volunteering information they have that IT will need addressed for SLAs: uptime and availability, user authentication, data access and retention, encryption, connectivity, 24/7 support options, and so forth. That means the responsibility still lies with the customer to establish the right SLA and to be proactive in ensuring they’re getting everything they need.

Business groups don’t always understand what should be included in a contract from a technology perspective, so partnering with IT early in the sales process will ensure the right SLA is established and subsequently enforced. One IT group I spoke with had established a brief questionnaire for cloud providers. Based on the responses, they would work with the business to make an assessment together. They also worked with legal and finance teams to create standard contract language that can be added to a cloud contract.

What sort of architecture should companies look for, especially for BI and data warehousing applications?

Self-service provisioning is a key value proposition of cloud computing, so multi-tenancy is critical. This shared infrastructure architecture creates economies of scale because all customers are on the same code line, usage can be easily tracked and monitored by the provider, upgrades are seamless, and new development, test, and production environments can be created with just a few clicks. It is generally easy to sign up online and get started with a trial.

Companies should determine up front what is involved in an implementation and ask how frequently upgrades and new releases are delivered. The answers should help to determine if it’s truly a multitenant architecture. When it comes to the architecture of a cloud data integration solution, it’s important to note that the service should not physically host customer data. Instead, there is typically a small server install where the processing takes place. Source-to-target mappings, transformations, workflows, and schedules are created online and processing is managed by the agent, which facilitates bi-directional data integration between systems.

What about performance concerns, especially with large data sets?

There’s a lot of talk about real-time cloud integration, as point-to-point application synchronization continues to be a primary use case. For example, if my sales team is using Salesforce CRM to manage accounts and opportunities, we also want up-to-date order and billing information to be available within the same customer record. Point-to-point application synchronization can do that.

There is also a tremendous amount of batch data integration that larger cloud application and platform customers must factor into their strategy. This is particularly relevant for the data replication use case. The right cloud data integration tool must be able to manage both types of requirements. A customer might begin with fairly tactical requirements, but inevitably, data volumes and complexity will grow with the company’s success. In that case, companies will want to avoid hitting the ceiling by using rogue technology that won’t perform and scale.

How is Informatica addressing the cloud computing opportunity and challenge?

Informatica laid out a vision for cloud data integration in 2006. Initially, we developed connectivity to leading SaaS providers like Salesforce.com. We realized quickly that the buyers, implementers, and administrators of these applications were not typically the same people we knew in IT, however. As a result, we developed a multi-tenant and on-demand cloud data integration service designed to address the needs of these users, called Informatica Cloud.

Today, more than 1,300 companies are running the service and it is processing over 80,000 jobs a day and integrating over 12 billion records a month. Informatica Cloud has now been recognized for three years in a row by Salesforce customers as the No. 1 data integration solution on the AppExchange. Trials are available at http://www.InformaticaCloud.com.