Andrea Danti - Fotolia
On Wednesday, during its virtual user conference, Tibco introduced Hyperconverged Analytics.
The vendor has long had a respected analytics platform. It suffered a short downturn in the immediate aftermath of Tibco's acquisition of Spotfire in 2007, but for the better part of a decade Spotfire, which became Tibco's main business intelligence suite, has been among the more innovative, helping lead the industry along with Tableau and Qlik when data visualizations ushered in a new era of business intelligence and continuing to push boundaries in the current era with augmented intelligence features like natural language processing.
But with all the tools vendors such as Tibco offer, a lot of maneuvering needs to be done to get to the point at which data is ready for analysis and ultimately fosters data-driven decisions.
Now, Tibco is bringing a host of capabilities together in one environment to create the Hyperconverged Analytics experience. The experience joins visual analytics, data science and streaming analytics in one location, and chief among the tools it takes advantage of are Spotfire 11, which was unveiled on Wednesday, and Cloud Data Streams, a new streaming data capture platform.
Nelson Petracek, CTO of Tibco, recently took time to talk about Hyperconverged Analytics and other new innovations the vendor, founded in 1997 and based in Palo Alto, Calif., planned to reveal during Tibco Now.
In addition, he discussed the trends he's seeing in the analytics market, how COVID-19 has affected innovation, and what Tibco has in the works for the next year.
How does Tibco define Hyperconverged Analytics?
Nelson Petracek: Tibco talks about itself with three different pillars. We have the connect portfolio, the predict portfolio and the Unify portfolio, and a certain set of capabilities fall under each umbrella.
Nelson PetracekCTO, Tibco
Hyperconverged Analytics falls under the predict portfolio. To us, the predict portfolio is very much focused around analytics -- this is visual analytics, streaming analytics, data science, artificial intelligence, machine learning, model ops, auto ML. Spotfire would be the product people would know best underneath that part of our portfolio. From a strategy standpoint, we want to remain very open as a platform. We want people to be able to plug and play, mix and match Tibco and non-Tibco components, and build their architecture up in whatever way they want. But we also had a lot of customers come to us though and say, 'We love that you're open and we can mix and match, but can you tell me what to do and be just a little more prescriptive?'
So now under the predict portfolio, Hyperconverged Analytics is this idea of allowing organizations to build analytical applications that are immersive and smart. They're very visual, they can be extended with new visuals, and they have an aspect of data science and an aspect of streaming and real time. It's the combination of a smart, immersive experience with real time or streaming side-by-side, and also including data science -- artificial intelligence, machine learning, data models. Hyperconverged Analytics is bringing those three capabilities together and allowing companies to build applications that use as little or as much of those features as they want.
What market forces was Tibco seeing that motivated the development of Hyperconverged Analytics?
Petracek: We saw a couple of different things. It's a trend that's being noted by industry analysts like the Gartners of the world and so on -- they have different papers around this blending of these worlds, and they have a term for it as well. But we also saw it within our customers. Our customers were saying that they wanted to have a view into their manufacturing floor, for example, but a historical view wasn't good enough. They want to see what's happening now, so they need that real-time view as well. We saw people wanting to blend in data science and make it more readily available to other personas, and then combine historical and real time and do all that in a single application. That's what we saw, and we had the three separate products and thought we should be able to come up with a framework and a product strategy and product features that enable organizations to do all that.
Tibco Cloud Data Streams is a new feature that plays a prominent role in Hyperconverged Analytics. What exactly is it?
Petracek: That is the streaming analytics engine that underpins real-time visualizations and real-time analysis. It gives you the ability to connect to a whole variety of streaming data sources -- anything from Kafka to OPCUA [Open Platform Communications Unified Architecture] and everything in between. You can use this engine to connect to streaming data sources, collect real-time data and then process or transform or aggregate or manipulate the real-time data streams -- join them, filter them -- and then forward that data to Spotfire for visualization in real time. Its focus is on the streaming part of data. Building an engine that's optimized for streaming data is very different than building an engine that's optimized for handling a million rows of data that I want to bring into my workspace and slice and dice my way through. When you put the Spotfire data manipulation engine and the streaming analytics engine together you get a single application that can combine both worlds.
Why is streaming data capture so important?
Petracek: People define streaming in different ways, and that's the first thing you have to address. Now, what you find is that the notion of streaming, because of the evolution or the increase in popularity of technologies like Kafka, can mean just high-speed data ingest into my data lake. I think that because of the renewed focus on data, where the data is coming from, the need to collect from more sources -- and things like IoT are just going to broaden that problem out even more -- streaming has become more interesting and a hot topic to people again. Data has become a first-class citizen for organizations, and streaming is one mechanism by which you get data into your data lake via things like Kafka, IoT and so on. It's kind of gone through these waves where it was kind of specialized, then it was more general but more for some verticals, and now it's a hot topic again to support an organization's data strategy.
Hyperconverged Analytics is being unveiled while we're still in the midst of the COVID-19 pandemic; how have product development and collaboration at Tibco been affected by people working from home rather than together in an office?
Petracek: It has definitely required an adjustment in terms of how we engage. There's the aspect of how you share information, there's the aspect of how you collaborate, there's the aspect of how you share thoughts and ideas and how you develop those thoughts and ideas. Previously, everybody would be herded into a boardroom in front of a whiteboard and you would have sorted things out. We've had to come up with new and different ways to drive the same conversations and reach the same outcomes. And I have to say, anybody that knows me knows that I'm the whiteboard guy. I'd fly everywhere, and if you gave me a marker I'd fill up five whiteboards in half an hour, so for me it was a bit of a shift. But we've derived a series of innovation processes to support innovation cycles supported by tools like Miro and ClickUp, some of these collaborative tools so everyone can see the same screen and collaborate on that screen.
We also found that people talk about being agile, but under this circumstance it's really important. We found that you can't get an idea and then go away for a week. A week is even too long. There has to be a more frequent cycle of interaction. It's been a combination of getting used to collaborating in a different manner, using some of these tools, and then changing the structure of our innovation process. I've seen some very creative use of virtual sticky notes where I've looked at a Miro board and there are 5,000 stickies across the board and now we have to make sense out of it all.
Have there been any frustrating moments?
Petracek: Yes. I wouldn't say it's all been smooth -- definitely not. It's us working through a new process, and it's also our customers working through new processes, and sometimes we're not on the same wavelength because perhaps there's been a certain assumption that might have been easier to identify in person -- you're missing that interaction, the ability to read facial expressions. There have been times when there may have been a mismatch on a goal or a mismatch on an outcome. So yeah, I wouldn't say it's been all smooth sailing. But if there's the right focus on team and collaboration you can work through those things.
Beyond the rise in streaming data, what are some other market trends Tibco is following?
Petracek: Some of the new and interesting things we're hearing from customers is this ability to push processing out closer to the edge [where the data is collected]. They can't bring all their data back to the cloud, they can't bring it all the way back to the data center and make a decision and then push the answer back because it takes too long. There's too much data, so they need to do more of that processing out at the edge.
Another interesting trend with that is the area of computer vision, the ability to process streams of video at the edge and then make decisions at the edge in order to identify a threat or an opportunity. That is becoming an interesting topic for a lot of companies. This whole IoT edge, AI/ML at the edge -- one of which is computer vision -- is getting a lot of interest and a lot of traction.
I find another one that's interesting is this idea of digital twin. You've probably heard of that in relation to a piece of equipment, but that notion is getting broadened where I want to mimic the way my organization runs its business and then be able to tweak that without having to go and change my organization. I need to figure out how to get the biggest bang for the buck, and to do that I want to create a digital twin of my company.
Another thing is AR/VR because everything is moving to remote, and the idea of being able to interact with your physical environment in a very digital and immersive way is getting a lot more interest. We get a lot more questions around that these days.
Now that you've unveiled Hyperconverged Analytics, what does Tibco have on its roadmap for the next year or so?
Petracek: It goes back to the three themes we have for this year. You've got the theme around any data, the theme around real time, and the theme around connected experiences. When you look at it from a roadmap and product strategy standpoint, that's what our focus is going to be around -- extending our ability to connect to more data, more sources, regardless of where it might sit, and then supporting things like the ability process data closer to where it's being generated out at the edge. That's going to be one part of it. To be able to do so at the speed at which it needs to be one is the real-time part, and Tibco has always been real time; that's how we got started. That is one of our core tenets, this ability to handle things in real time.
The last one, connected experiences, is where you're going to see a fair amount of work. This is the ability for us to offer a more seamless product experience for our customers, and this is not just within a single product but across products, so how can we make it easier for people to consume different product capabilities without making it seem like it's different products? That's going to be a big focus of ours. We see more organizations wanting to build applications that contain bits and pieces of all these different domains -- a little bit of analytics, a little bit of APIs, and little but of event processing, a little bit of data science -- and they want to create and expose all through a single application. Because we see that trend, our products have to further evolve to support that trend.
Editor's note: This Q&A has been edited for clarity and conciseness.