Why Companies Still Choose On Premise Business Intelligence Deployments
By Editorial Coordinator for TechnologyAdvice.com, Cameron Graham
The growing popularity of the cloud
These days it’s difficult to have a conversation about the future of technology without mentioning cloud computing. For invested companies, such as Microsoft and Amazon, cloud computing is the future of business technology. Whether by accelerating innovation, enabling new business solutions, speeding time to market, or just promoting cost savings, these companies believe it’s no longer a question of if, but when, most companies will adopt cloud technologies.
In many ways, Business Intelligence software seems like a natural fit for the new cloud-based office. Analytics solutions, like Yellowfin, are designed to process incoming data in near real-time. Increasingly, these data sources are both generated and housed online. Executives also want the ability to view (and interact with) their BI dashboards from anywhere in the world. Software as a Service (SaaS) programs offer them this flexibility, without the need to jump through extra IT hoops.
Yet, most enterprise BI companies still offer on premise deployments. Largely, this comes down to two factors – data transfer speeds and security. Solutions are beginning to emerge for both problems, but they remain a hurdle for wary enterprises.
Cloud-based Business Intelligence concerns: Data transfer speeds & security
The growth of cloud computing has unlocked mountains of previously inaccessible data, from smart phone user behavior, to the data from internet connected appliances. Yet, in order to efficiently process this data, cloud computing has to overcome its own technological limitations. The problem with moving enterprise-level volumes of Big Data in the cloud is that they’re … well, big. Sometimes as big as exabytes, or one billion gigabytes.
Uploading this amount of information from an on premise network to the cloud over a standard connection is like trying to suck up all the water in Lake Michigan with a household vacuum cleaner – in other words, highly ineffective. There are a few software vendors, such as Aryaka, that are developing solutions to help optimize wide area network (WAN) connections to speed up such data transfers. However, this type of solution is a work in progress and still requires a large time investment.
Jeff Kelly, the Principal Research Contributor for professional tech community Wikibon, stated at BigData SV 2014 that “it’s going to take some time before the cloud becomes a place where you are bringing your enterprise workloads.”
That’s why enterprises which produce most of their data in-house (such as financial companies), or have extensive private-cloud infrastructure already in place, generally opt for an on premise deployment.
The second issue is security. When building a private cloud, companies purchase the servers and infrastructure up-front. While these investments come with maintenance costs, they also allow companies to have direct control over their servers and security systems. While physical installations are certainly not immune to hackers – see Target’s recent data breach – it’s a lot easier for companies to restrict access to their own hardware than a server located across the country, or overseas.
Public cloud platforms: Risky business?
Enterprises are becoming more open to the idea of public clouds – a recent Verizon survey found more companies than ever are putting data on public cloud platforms – but recent security lapses such as the Heartbleed bug are likely to be a setback. And for some, then there’s the continued wariness over the US government’s digital surveillance initiatives.
These are the concerns that enterprises have to consider when deciding between a SaaS BI solution, and an on premise deployment. Luckily, most large-scale BI companies offer both options.