Strong Growth And Innovation Seen For Information-As-A-Service In 2011

Over the course of this year, I’ve spoken with many organizations that are continuing to expand their usage of information-as-a-service (sometimes called data services) to support new business requirements such as self-service customer portals, real-time BI, and single-version-of-the-truth. With the growing complexity of data, increasing volume of data, and exploding security challenges all driving demand, IaaS is poised to grow significantly in the coming years, especially as existing integration technologies are failing to meet these new requirements. What we see is that most organizations that have embraced an IaaS strategy over the years aren’t looking back; they’re continuing to expand its usage to support more requirements such as real-time data, creating data domains, improving the ability to securely deliver information, integration with unstructured data and external sources, various Web portals, and enterprise search.

Recently, my colleague Gene Leganza, who serves Enterprise Architecture Professionals, compiled the top 15 technology trends EA should watch over the next three years . One of the trends Gene highlighted is that information-as-a-service (IaaS) is finding a broader audience. I see more organizations continuing to show strong interest in IaaS, as evidenced by increasing inquiries, to help with growing data integration challenges that traditional solutions are not addressing. IaaS can significantly alter IT’s approach to its data management strategy and delivers a flexible framework to support transactional, BI, and real-time data.

Here are my top predictions for 2011 related to IaaS:

  1. Large and complex IaaS deployments will become more visible. Some of the largest and most complex IaaS deployments supporting tens of thousands of users with terabytes of information will become more visible in 2011 and beyond and span various vertical industries, including telecommunications, government, retail, and financial services. In addition, many will become more mission-critical to support global applications and new user requirements, requiring redundant infrastructure and processes. 
  2. More organizations will focus on near real-time information. Most IaaS deployments we have seen so far tend to focus on refreshing data every day or every few hours, but we definitely will see organizations ramping up the frequency to support real-time or near-real-time data in 2011 and beyond, especially as the business value of IaaS becomes more visible across the enterprise. As a result, IaaS going forward will become the preferred source for trusted data in the enterprise. In addition, organizations will look at integrating IaaS with other real-time sources including streaming data and CEP systems. 
  3. More will use data services (IaaS) to speed mobile apps delivery. I recently heard about a large insurance company, a Forrester client, that attributes its success in rapidly rolling out new iPhone and iPad apps in part to the foundation it had laid with SOA. This included both transactional services, which can be triggered by iOS apps (and others), and data services, which securely and rapidly deliver trusted information to mobile devices. This is a great example of how to leverage your investment in SOA to help you meet the exploding demand for enterprise mobile apps. 
  4. More data sources will be added beyond the typical four or five. So far we have seen enterprises integrating, on average, about four to five data sources to support IaaS. This is likely to increase to tens and in some cases hundreds as more enterprises start to leverage IaaS in a strategic manner. In addition, we estimate that in 2011 and beyond, half of all IaaS projects will have a more enterprisewide — as opposed to project-level — focus. 
  5. Demand will increase for integrating structured with unstructured data. We find that more organizations are realizing the importance of using unstructured data to support their decisions. As a result, integrating structured with unstructured and semi-structured data to support new and complex business decisions is becoming critical. In 2011 and beyond, more organizations will start to look at solutions that can deliver a simplified integration process of structured with unstructured data. 
  6. More organizations will realize IaaS’s security potential. During various inquiry calls this year, some customers mentioned to us that they dramatically improved data security by using an IaaS framework, centralizing the authentication, authorization, and access control of critical data — as a result disallowing direct access to data sources. This not only improves compliance audits for various organizations but also ensures that only authorized users can access sensitive data. We expect that more organizations, especially those that have hundreds and thousands of sources, will adopt an IaaS framework just to improve data security and compliance. 
  7. Read-write information services will become more common. Most organizations that we spoke with during this year have read-only IaaS deployments, with a few read-and-write, but we believe that will change in 2011 and going forward. We find that typically organizations try out read-only information services first and then over time leverage read-write capabilities, especially when they make IaaS a more strategic platform in their organization. 
  8. Even smaller and moderately sized companies will start to look at IaaS. So far, most of the IaaS deployments have been concentrated around very large $1B+ companies; however, we are now seeing that even smaller and moderate sized organizations are very interested in IaaS to help them overcome their data challenges. As a result, in 2011 and beyond we are likely to see small mission-critical IaaS deployments in these organizations to support various types of use cases, ranging from single-version-of-the-truth to BI to searching and compliance reporting. 
  9. More will use IaaS to enable integration with external data sources. Some large organizations have already started to integrate their on-premise data with partner data or external sources such as the public cloud to deliver more value to applications and end users. Microsoft recently announced Windows Azure Marketplace DataMarket, which delivers data from commercial data providers and public data sources, such as data sets on demographics, financials, retail, weather, and sports. We are likely to see at least a fourth of all IaaS implementations integrating with external or public data sources in 2011 and beyond.

 

Comments

Your blog post on Iaas

Good insights in the post. Your blog post is music to our ears and to many companies like us that are trying to bring unstructured data not just from inside the firewall but from the outside world into the decision making.

Considering Infrastructure as a service has already taken IAAS acronym, we may have to invent a newer one so as not to confuse with that. Possible candidates are DAAS (data as a service), BAAS (business intelligence as a service). May be we can restart the trend of TLAs and use IAS (Information as Service), BIS (business intelligence service). Well, I am sure you and your team at Forrester is expert at creating new trends.

Best Practices for Data Virtualization

Hi Noel,

Great post.

The terms "data services" and "data virtualization" seem to come up as alternatives when I discuss IaaS or Information as a Service with end users and architects.

Also, architects almost always look to base their architecture on the best practices recommended by the Forrester IaaS Wave and the work you and Mike did on the "information fabric."

In the numerous discussions I have had with architects on data virtualization and data services, one thing that stands out is the end goal - "agility." They believe that "agility" is all about cutting the "wait" and the "waste" involved in a typical data integration process. As we know, simple, traditional data federation is not data virtualization - although the lines are blurring across the web, creating some degree of confusion.

Data virtualization should be all about hiding and handling complexity. Enterprise data is complex as we all know. It requires rich transformations not limited to SQL or XQuery only. Data federation assumes data is ready for consumption and of good quality, which is not the reality. You can't propagate bad data in real-time and waste time with post-processing. Data profiling, data transformation and data quality (including data masking) must happen on the fly, on federated data. And this is not easy. Additionally, you cannot leave any data behind, simply because you can't access it (high performance), specifically in the case of unstructured data and mainframe. All this can remove the "waste" in a data integration process. Finally, business-user self-service or instant business-IT collaboration using role based tools can help remove the "wait" from the data integration process.

We at Informatica, a leader in the Forrester IaaS Wave, have put together a graphic based on best practices for data virtualization and data services, that we are hearing from architects - it can be found here:

http://www.informatica.com/INFA_Resources/ds_data_services_7010.pdf

http://www.informatica.com/products_services/data_services/Pages/index.aspx

http://www.informatica.com/Pages/data_virtualization_index.aspx

Thanks.

Ash Parikh