Last week I got a chance to attend the 11th Annual E-Business Conference that was put on by the University of Wisconsin E-Business Consortium. (www.uwebc.org) It was a great event and well attended, given the economic conditions.
One of the speakers was Kevin Leahy, who is the Director of Strategy & Marketing for IBM Global Services. He spoke specifically on the advances and challenges of cloud computing. In his presentation, he raised an issue with cloud computing that seems to be a root cause for slower than expected adoption of cloud computing for business intelligence. Seems that the key culprit is data transfer times.
Full disclosure here…I am not a networking expert…and I didn’t sleep at a Holiday Inn Express last night. (It was a Marriott!) However, it seems to me, that even some of the fastest Internet connections still cannot compare to Gigabit Ethernet connecting multiple servers on the same switch. Therefore, data volumes will clearly need to be part of the equation when making the decision to move BI to the cloud. Another variable in the equation will certainly be the refresh frequency. If you need large amounts of data updated frequently (maybe several times per day) BI in the cloud may become even more challenging.
For many companies, BI in the cloud gets stuck in the mud right out of the gates because they see how much data has to be moved just to get going. The initial transfer seems a bit scary. Then they start thinking about all their data passing through the Internet and they get even more nervous. Another option, often recommended by some of the cloud computing giants, is to ship your data on disks for faster loading. Yep…good old-fashioned disks. As a former IBM employee, I can tell you just how risk-free data transfers on disks can be. I got a free year of credit monitoring from IBM when they lost a disk that had my SSN on it as a former employee! Maybe that is not the best option.
Now there are some emerging tools to help make BI in the cloud a better and more realistic option. Google and others are starting to offer temporary cloud-based VPN solutions to safely move data over the web. Other companies like Freedom OSS and NetEx are starting to offer solutions to make it faster to move data over the web. There is still the issue surrounding the BI vendor you choose that is offering the cloud-based BI solution. LucidEra is now the poster-child of failed, SaaS-only BI vendors.
So what should a company do? Well, clearly BI in the cloud is not going away. The concept should be included in your BI strategy at some point. Depending on your IT infrastructure, data volumes, data refresh rates, etc. you might find BI in the cloud makes better sense in the near term. However, keep in mind the issues mentioned above where others have struggled and head to the cloud with your eyes wide-open.
Until next time…Shawn