Many (if not most) companies are evaluating the benefits and risks of cloud-based solutions this year. In fact, marketing research firm IDC predicts that businesses will spend $22.6 billion on cloud services by 2015. However, there is one area that has fallen behind the cloud – business intelligence. But it's ready to emerge. Even organizations with traditional (hosted on-premise) BI systems in place can make the move. Let's consider the practicalities of doing so.
Organizations that have deployed business intelligence have first-hand knowledge of the complexities of such a system – the vast network of linked parts and pieces, from data warehouses to ETL applications, OLAP servers to analytical dashboards. It's a jungle out there and it's clear that it can't continue this way for much longer. A more repeatable and sustainable model for business intelligence must emerge – one that reduces the complexity while maintaining security and enhancing ease of use.
The Data Question
For services like CRM and document collaboration, the roadmap for moving to the cloud has already been established by companies like Salesforce.com and Google. But for BI, it's not as clear. The sensitivity and volume of data as well as the inherent complexity of BI systems have left executing a cloud-based BI strategy more of a dream than a reality.
Many believe that the next logical step in BI's evolution is moving it to the cloud. However, when looking at the characteristics of a modern day BI deployment, it's easy see how getting there is complicated.
Let's take a look at just one aspect of a cloud BI deployment: the amount of data that would need to be moved, stored, and processed. There's a reason we're all talking about big data these days – according to April Adams, research director at Gartner, data capacity in enterprises is growing at 40% to 60% year over year. We're not just looking at the past 12 months of data – we want 3+ years for longer-tail analysis, so we've got to store that backlog. It's not unusual for a large organization to have information equaling tens of billions of rows, which doesn't even account for all of the unstructured data brought in from technical support systems, CRM software, and social media.
Even a small BI system produces deluges of data from a relatively small set of users. Likewise, at small to medium-sized businesses, there are potentially thousands of transactions being processed daily. BI systems take transactional data and use that as a basis to create more data, resulting in the data warehousing and data mart infrastructure we see today. So the question becomes, is it going to be cheaper, faster, and easier to move all of this data to the cloud? For some, yes; for others no – but you don't necessarily have to move your data to the cloud in order for your BI to be there. No, really.
Getting BI in the Cloud Securely
The creation of "on demand" web services can now circumvent the requirements for significant infrastructure and data extraction, transformation and loading. This means that existing data warehouse and data mart infrastructures can be securely leveraged from within the corporate firewall by utilizing web services-based SOA broadcast services.
These broadcast services are secure, standard BI queries that have been tuned behind the corporate firewall to optimally perform in the cloud and be consumed by BI and analytical systems. They can be secured using robust Internet encryption protocols, mitigating security issues, and can be incorporated seamlessly into desktop client or mobile device BI consumption mechanisms.
Gartner predicts that this year, 70% of global 1,000 companies will host BI in the cloud. If you're considering this, BI software-as-a-service (SaaS) and web services might be a viable option for your company if data security is paramount, or if migrating the amount of data you're storing would be too difficult.
Next time, I'll evaluate the risks vs. the rewards of migrating your existing BI into the cloud.