In this interview, learn why business intelligence cloud solutions are here to stay, but not as you necessarily expect.
It’s up there. All your information. It’s floating around us. Or at least that’s what most of us think when we hear about cloud solutions. Everyone’s jumping on the bandwagon. Cloud services are popping up all over the place. But is it just a fad, a passing phase, like the MiniDisc? Or is it here to stay? And how is it transforming the world of BI?
To get the inside scoop, I chatted with our very own Head of Product and BI guru Dr. Guy Levy-Yurista.
A: If you open up Wikipedia (trusted to a degree, for sure) , you’ll see cloud computing defined as « a type of Internet-based computing that provides shared computer processing resources and data to computers and other devices on demand. »
In other words, it’s a way to outsource your IT headaches to someone else, for a fee. This « someone » will handle your entire IT infrastructure at their end, giving you more flexibility; the ability to increase or decrease the number of resources you’re utilizing in a very flexible manner.
A: When there’s a new technology fad, everyone hops on the bandwagon and tries to figure out how to apply it to their field. The cost of missing out on a major cost saving, groundbreaking technological innovation is too much to bear. The same happened with cloud technology in industries across the board.
And at some point, someone said, « Everyone is talking about the cloud. Let’s put BI in the cloud and claim victory. » So companies started offering pure BI solutions in the cloud, where customers’ data would be stored on a big machine somewhere, sitting in a data center — like, for example, AWS or Azure.
This created a multitenancy environment where many customers were running their BI and Analytics (BI&A) implementation on the same application running on the same cloud machine, and all the data flowed to that same machine and application. Pretty quickly, though, people realized that there were several limitations to the so-called « pure cloud, » most noticeable of these limitations have to do with the notion of data gravity.
You see, massive amounts of data are used in BI&A. And the more data there is, the closer you need to be to that data, computation-wise. You need to upload the raw data to the cloud, download the processed data from the cloud, and so on. So there is a massive amount of traffic when it comes to BI cloud solutions, which is also often associated with hidden costs. There are also other hidden costs that have to do with the specific architectural structure of AWS/Azure/cloud-provider-of-choice but I’ll skip those for now (you’re welcome) .
Cloud is now headed into a hybrid mode, where you have a combination of on-prem data, data-stores and a cloud platform – and you can easily switch between these situations. Essentially, hybrid cloud solutions alleviate some IT concerns, but don’t remove IT completely. At the same time, it doesn’t require customers to upload their entire data – which is often very sensitive data – into a cloud environment.
A: In short, a hybrid cloud is definitely a cloud. But it is a cloud offering with distributed resources; some of them are on-prem on the customer side and others are on the cloud. Think of it as the best of both worlds.
A: So data gravity is the major one. But other disadvantages have to do with PII (personally identifiable information) , PHI (Protected Health Information) , PCI DSS (Payment Card Industry Data Security Standard) or other costly identifiable information and sensitive data that you don’t want to get off premise.
Or, for example, if you have an EU entity, then you are not allowed to take certain kinds of this data out of the EU (see General Data Protection Regulation (GDPR) for more info) . So there are now a lot of laws and regulations, globally, that limit your ability to use a pure cloud solution or data centers located remotely.
Now if you are a cloud provider and you don’t want to establish cloud or storage presence in the EU, then you can still leverage the local storage presence of a customer in the EU, but at the same time have a computation presence somewhere else.
So there are a lot of ways to keep in line with laws and regulations and still provide the huge advantages of cloud technology.
A: Realistically, the prime advantage of the cloud is that it allows you to rapidly scale up or down services so that you can get your offering out to vast quantities of people very, very quickly, and at the same time react quickly and efficiently to changing market demands. Let’s say you’ve had a whole bunch of new customers sign up and now need to provide them with, for example, analytics. With a cloud solution, you don’t need to spin up new machines or put IT to work. You simply take a machine and clone it in the cloud. It’s a one click experience.
Networking is easy. People like AWS or Azure provide you with the ability to automatically scale up and down resources as needed. So there are a lot of operational effectiveness and efficiency gains to using cloud.
A: One way to ensure security is to build a cloud instance dedicated to one customer only and then let that customer manage this instance so that no one else can touch the data. There are now advanced certificate and encryption mechanisms that have been put in place to ensure this outcome. Then the BI service provider can play around with the IT or the network, but it cannot touch the data. This way the data is fully under the control of the customer, but the benefits of cloud scalability and efficiency are still at play.
A: First of all, they should figure out why they want to use a BI cloud solution. Is it just because they’ve been hearing this buzzword all over the shop or is it because of a real need? If there is a real need — for example, efficiency reasons — then that’s great. But they need to get familiar with the security and privacy concerns, as well as the regulations they need to meet in their own sector, environment, or geography.
Then they need to plan accordingly for a pure cloud or hybrid cloud solution and understand the different levels of attention to sensitivities needed. Do they want a managed service where someone else is managing their machines for them? Or do they want to do a BYOL (Bring Your Own License) , where they buy the license and then just get machines while they manage the cloud instance completely on their end?
They also need to figure out if they want a multi-tenant service, where they don’t mind sharing the same machine with other companies — or if they want a dedicated instance service.
