What Challenges Arise With BI on the Cloud?

One needs to first understand what customers are expecting from a BI product in the cloud, is it Operational BI - which is more transactional; or Traditional BI - which is more strategic in nature.

With the current economic climate, I believe that Operational BI would be what customers will be interested in getting a handle on their immediate business concerns.

From that perspective the challenges would be:

• Common data exchange mechanism between the application providers in the cloud


• Data Security from the customers view point as sensitive data could be compromised unless it’s on a private cloud


• Customization of Key Performance Indicator (KPI) libraries which customers would want to define - as every organization would have their typical business/org specific KPIs


• Scalability of BAM/BPM/CEP engine to handle high-volume transactional data.


• Storing and processing large volumes of data.


• Integration between data available on the cloud vs Customer provided data(On premise)

One of the biggest complaints and problems with current BI is the interactivity and the ability to change analyses on the fly. For example, if I wanted to do Region wise and then product wise analysis, the numbers will look one way. If I do product wise and then region wise then the numbers may need to be computed again or precomputed and stored for quick reference.

The tug of war between what is done locally and what is done in the cloud will be a tricky problem. Local computing is fast but data may need to be downloaded in sections and computation done. If the computations were remotely in teh cloud then it may be slower and the local computing is wasted just waiting for a browser response.

Interesting problem but someone may come up with clever answers!

There's no shortage of potential "areas of concern" with cloud-based BI, but isn't this also true of internally-managed information systems as well?

Areas of Concern:
1) Lack of control of "information ownership" or "data ownership" -- Ask yourself this: do you really have actual "control" of your data in either environment? ...Or is it truly less a matter of control than one of agreement-to-comply with some set of mutually-agreed-upon standards? Ultimately, whether cloud-based or internal, it comes down to knowing which behaviors are acceptable, and having an agreed-upon method of enforcing whatever usage boundaries you've come to expect from both users and providers-of-services. It won't matter much WHICH environment was used; if there's a serious problem, more often than not, it's going to come down to something other than the technology.

2) Lack of "awareness" of either security rules, business practices or technological limitations of a given BI environment -- if the expected rules and boundaries are not clearly communicated, agreed-to and actually understood, you can expect a rather wild degree of variation in compliance with any security, business or technology implementation. If it's a matter of a user or provider understanding -- but not agreeing to comply with -- a given set of guidelines, it's really not important what type of environment is being implemented, as much as whether or not that specific behavior is acceptable in your BI- and management-culture. Governance is key, and will need to happen OUTSIDE I.T., within the Business side of the organization, with IT as an "enabler," rather than the "owner" of the overall BI solution environment.

3) Sustainability has a cost. Well-designed business intelligence environments need to account for data quality, data integrity, and other fundamental characteristics of acceptable information-quality levels. While cloud-based BI might bring "new" capabilities in terms of connectivity and availability, it's certainly the case that those capabilities come with some costs-to-support. Most likely, those costs won't be well-understood up front, but will become much more clearly defined as the users evolve their needs, and the providers of BI-in-the-cloud services evolve their capabilities to meet those needs.

Suggestion: Keep track of both cost and business value, so you can judge the overall worth of the balance between cloud-based and non-cloud-based solutions in terms of what's important to your business. There's more to the equation than dollars: it's also important to actually measure how well you've met your sponsor's stated business objectives, some of which are NOT about cost, but are more about achieving organizationally-prioritized values.

Cloud-based business intelligence will solve a number of important, previously poorly-served areas of reporting and analysis. We, at Jaspersoft, have studied these areas (use-cases) during the course of our product planning and the launch of our cloud-based BI service, which is offered as an integrated stack along with RightScale, Vertica and Talend. This service offers "instant-on" and pay-as-you-go BI services that we believe will provide an elegant BI answer in at least four scenarios:

1. Enterprise Proof-of-Concepts where sequestered data and non-production BI systems are necessary.
2. Enterprise Projects or "Situational BI" where an event or project milestone is relatively short-term and for which a costly BI platform installation is not possible.
3. Small and Mid-Sized businesses that have neither the expertise nor the capital budget to install and then maintain a complete BI stack.
4. Other Cloud-based applications that require fast, rich, affordable analysis (either operational BI or strategic BI) and would want BI platform similarity.

Our experience with all four of these scenarios tell us that they share the need a for fast, low cost infrastructure, pay-as-you-go BI solution. The primary questions or challenges we hear about include: data security, integrated procurement and billing (for administrative simplification), and ability to customize. Future versions of our integrated BI stack in the cloud will address these primary challenges. The main lesson we learned through this product launch: BI in the cloud is of great interest to a sizable market and it will become mainstream more quickly than most expect.

Brian Gentile
Chief Executive Officer
Jaspersoft