SaaS Business Intelligence Checklist – Architecture and Features
This post is part 3 of a ten part series on the important things to consider when working with SaaS business Intelligence vendors.
Today’s topic is Architecture and Feature Sets for SaaS Business Intelligence Solutions
Cloud computing and software as a service (SaaS) both took significant steps forward in 2010. Early entrants to the space focused on delivering the basics in both features and functionality adhering to the 80/20 rule of providing early adopters most of what they needed. 2010 brought a new level of maturity to SaaS business intelligence and reduced the 80/20 gap. A growing end user community fueled the evolution by demanding stronger feature sets, more sophisticated integration solutions and more powerful architectures.
Some providers are coupling their SaaS BI applications with strong platform architecture (PaaS) offerings designed to scale with enterprise level clients and provide development environments. Recent research by EMA shows that 75% of those planning cloud projects intend to deploy private cloud environments versus Public. Utilizing a vendor that can deliver the platform as well as the application may enable you to drive your projects faster and further.
For end users looking to add SaaS to their business intelligence strategy, review of feature sets is extremely important. Its critical that you are diligent when identify the present day needs of your company and explore the scope of capabilities that will serve you in the future. Be sure to research the technology road map of perspective vendors so you fully understand where they are heading and how they plan to get there.
Most vendors are delivering SaaS BI in a multi-tenant configuration this means you will be using a shared code base as well as a single instance of the application this design enables SaaS vendors to make changes quickly and be more responsive with service and administration needs thus lowering costs. The down side to multi-tenant is the environment may prove to be inflexible and in some cases unable to serve your customization needs. Custom coding in a SaaS environment can be very costly, its critical that you explore the economics involved to customize a SaaS infrastructure before committing to a vendor.
Critical things to remember -
- Make sure the vendor has what you need now and is heading in the same direction as your needs
- Explore your customization needs upfront
- Vendors with packaged applications can jump start your adoption of SaaS BI
- Explore options that couple applications to strong PaaS solutions they may provide flexibility

Posted in Business Intelligence, Cloud Computing, Data Integration, SaaS Business Intelligence Tags: Business intelligence, Cloud Computing, Platform as a service, Software as a service

Platform architectures and application architectures are so fundamentally different that there is no way one product can be delivered as platform and also as an application (unless the application is “cloudwashed” into cloud platform). There is no packaged version of Salesforce.com or Google App Engine – these product are not designed to function as apps but as platforms and it requires completely different deployment, scaling and management.
Multitenancy doesn’t necessary mean limitation in flexibility. 100,000 of companies modified their version of SFDC and hundreds of companies use GoodData and most of them modified our product to fit their usecases. And we host only one version of GoodData platform – all customer changes are done at the metadata level.
Pingback: BI SaaS Checklist Part 3: 架构与特性