Mission

Search

Time Plus Data Equals Efficiency with Paul Dix, the Founder and CTO of InfluxData and the Creator of InfluxDB

Play episode

 If the topic of databases is brought up to certain people, their eyes may gloss over. But if that happened, that would be because they just don’t know the awesome power of databases. Data can be valuable but only if it is contextualized, and time is an extremely relevant aspect to consider when analyzing huge amounts of data. Paul Dix, the Founder and CTO of InfluxData and the Creator of InfluxDB, explains how a time series database can help provide that temporal contextual information to promote efficiencies.

Main Takeaways

  • Time Contextualizes Data: Data has value only when it is placed in context and then the information gleaned from it is applied into actionable items. Time is a key factor to provide a basis for understanding information. A time series database, like InfluxDB, can provide this sort of context for server and IoT device monitoring. This info can then be applied to track performance and increase efficiency.
  • Failure Becomes Opportunity: Sometimes it’s hard to see how a win can come from a loss, and most people try their very best to avoid losing. But the reality is that learning is happening when something is being created, and the knowledge that’s gained in the creative process has nothing to do with the outcome of a given project. To ultimately be successful, the lesson is to take what’s been learned and then keep pivoting until the product and the market are aligned and the timing is right.
  • Evolving Engineering: Technology is always advancing rapidly. Therefore, even a successful product will require adaptations to meet new challenges. Accepting the reality of the high rate of change and, therefore, the need to constantly adjust accordingly will position a company in the best position to succeed. 

For a more in-depth look at this episode, check out the article below.

Article:

If the topic of databases is brought up to certain people, their eyes may gloss over. But if that happened, that would be because they just don’t know the awesome power of databases. Data can be valuable but only if it is contextualized, and time is an extremely relevant aspect to consider when analyzing huge amounts of data. Paul Dix, the Founder and CTO of InfluxData and the Creator of InfluxDB, explained how a time series database can help provide that temporal contextual information to promote efficiencies.

“​​Collecting all of this data is basically so that you can make more intelligent systems and you can optimize processes,” Dix said. “I know of some use cases in factories where they want to optimize the way the machines run for power consumption. Because when you have say like 50 factories around the world and they’re running all the time, reducing your power consumption by 10% by optimizing the way your machines run is a huge literal cost savings. So a lot of the times that’s what you’re doing is you’re optimizing some function: a way a machine runs. Predictive maintenance is another big use case. So you can say, ‘Oh, we know this machine is going to fail so we’ll take it out of service first and fix it or do a repair before that happens.’ For example, Rolls-Royce Power Systems uses Influx for just this kind of thing.”

Due to the cloud, there is the ability to hold amazing amounts of information. Additionally, there are sensors that have the ability to take in massive amounts of data from specific vantage points. What’s needed are ways to understand all of this data so decisions to drive productivity can be made accordingly.  

On a recent episode of IT Visionaries, Dix shared how his company has created a time series database, called InfluxDB, that provides this time-based, contextualized data. He also chatted about the process that led him to co-found his company, InfluxData. Additionally, Dix described how the engineering work on his database, InfluxDB, is always ongoing because the database must adapt to constant technological advancements. 

Dix defined that there are actually two kinds of time series databases.

“There are two different kinds of time series data,” Dix said. “There’s what’s called regular time series, which is what most people think of when they think of time series data. That’s like metrics. So that samples [are] taken at fixed intervals of time; like once a second, once every 10 seconds, once an hour, or once a day. In some sensor data use cases, you may actually be taking measurements like a thousand times a second. But then there’s irregular time series data, which is essentially event driven. That’s just tracking events as they occur. So that could be individual requests into an API, [or] that could be a machine turning on or off [or] a state changing somewhere or something like that.” 

Like many founders, Dix has had an interesting career trajectory where he learned different lessons along the way that ultimately led to co-founding InfluxData and creating InfluxDB. At first, he started building a system for a fin tech startup that had to chronicle data based upon intervals of time. Later, a SAS product Dix worked on, that ultimately didn’t succeed, is what inspired him to build the time series database product that became InfluxDB.

“I initially started this company actually as a company called Errplane, and that product was essentially going to be a SAS application for doing real-time metrics and server monitoring and stuff like that,” Dix said. “So that was similar to like what Datadog is or New Relic or Stackdriver, those kinds of things. To build that application, so this is in a completely different problem, domain, [and] a completely different space, I used the exact same technologies. I basically had to build a time series solution to store that kind of data. So that’s what gave me the idea that time series was actually just a useful abstraction for solving problems in a number of different domains.”

For Dix, InfluxDB was born from this experience and then the product took off. He bet that there would be many areas where the database could be used and this has been proven true.

“Server monitoring is a big use case,” Dix said. “[This includes] tracking what’s going on in your servers and your applications [and] application performance monitoring [and] network monitoring. But sensor data is more and more a bigger use case for us. And that could be industrial IoT tracking what’s going on in a factory or in a warehouse or in solar farms or power plants or any of that kind of stuff. But it also could be consumer stuff.” 

Not only is sensor data becoming a larger part of the way InfluxDB is used. Dix is also very bullish on the innovation that’s being driven in this area.

“I get mostly excited about the sensor data use cases because they’re just so interesting and they’re so widely varied, ” Dix said. “One sensor data use case [involves] people growing apples for hard cider that they’re making. And they’re putting sensors into the apple trees to basically optimize their apple production… Another is SunPower is using InfluxDB to monitor energy production and sensor health on solar panels [that are] powering homes… Basically it tells me that literally anything that can be instrumented out there in the physical world will be eventually. So just think about the limitless number of sensors that are going to be installed all over the world [and] taking measurements — all different kinds of measurements.”

It could be reasonable to hope that once a great product is made, then the work is done. But the reality is that as technology accelerates so does the need for new product versions.

“There’s always more data that people could be collecting, so they could get better visibility into their systems and their sensors,” Dix said. “So we’re in the process right now of basically creating a new core of the database that will enable us to have essentially unlimited cardinality [and] unlimited number of series scaling up to multiple petabytes of data across tens of thousands of servers. Basically, for every engineering challenge, you’re lucky if your solution will last for two orders of magnitude of scale. Likely, you only get one and essentially you end up having to redo the whole thing to get the next order of magnitude and then the next one.”

Certainly, the potential for contextualized data being used to drive innovations is very promising. The fact that InfluxData is hard at work adapting InfluxDB to meet whatever amount of data is captured is even more encouraging.

To hear more about how Infux Data, through its database InfluxDB, is providing temporal context to data that can be used to help productivity, check out the full episode of IT Visionaries

IT Visionaries is brought to you by the Salesforce Platform – the #1 cloud platform for digital transformation of every experience. Build connected experiences, empower every employee, and deliver continuous innovation – with the customer at the center of everything you do. Learn more at salesforce.com/platform

Menu

Episode 330