InfluxData raises $81M to advance time series capabilities
The vendor is the creator and lead sponsor of the open source InfluxDB database and plans to use the new funding to further product development as it aims for profitability.
InfluxData, a specialist in time series data management, has raised $81 million in a series E capital funding round and other corporate financing to bring its total funding to more than $200 million.
The new funding, revealed on Feb. 8, includes $51 million from venture capital investors, increasing InfluxData's total VC funding to $171 million, plus an additional $30 million debt facility with Silicon Valley Bank.
Founded in 2012 and based in San Francisco, InfluxData is the creator and lead sponsor of InfluxDB, an open source database purpose-built to manage time series data and enable time series analysis.
Time series data is a collection of data gathered at successive intervals and recorded in the order the events captured took place. Time series databases, meanwhile, are systems optimized to sort and organize data based on the time the events captured took place.
In addition to InfluxData, vendors such as Grafana and Prometheus offer time series databases. Tech giants Amazon, Google, IBM and Microsoft also now provide dedicated time series platforms.
Beyond InfluxDB, InfluxData develops projects that complement the database, including the Flux query language and Telegraf metrics agent.
Recently, InfluxData CEO Evan Kaplan discussed the vendor's latest funding round, including what nearly doubling its total funding will enable InfluxData to do that it previously could not and some of the plans for how the vendor will use the new money.
In addition, he spoke about the importance of time series analysis and why it's a growing aspect of data and analytics.
Princeville Capital and Citi Ventures, both new investors in InfluxData, led the series E round with participation from existing investors including Battery Ventures, Mayfield and Sapphire Ventures.
What is time series data?
Evan Kaplan: It's actually pretty conceptually easy to grasp. Think about any measurement that is timestamped -- the universe of those things is tremendous. Anytime you're instrumenting something -- for example, in baseball it could be measuring the velocity of the ball as it leaves the pitcher's hand until it gets to the catcher's glove -- there are time-based measurements. Anything that is a time-based measurement that can be captured and recorded and then be analyzed and used to predict future behavior becomes a kind of time series data.
What is the advantage of a database dedicated to time series data versus a typical database that handles all types of data?
Kaplan: Typically, this type of data used to be handled by a general-purpose database. If this were 20 years ago, someone might have built something in Oracle's relational database, for example. If this were 10 years ago, someone might have used a general-purpose SQL database. But over the last five to seven years, there's been an emergence of this category called time series, which is that if you know the shape of the data -- which is measurements over time -- then you can optimize a data platform for four or five things that really matter.
What are some of those things that make a database optimized for time series?
Kaplan: One is how quickly you can write data. Imagine you're dealing with 100,000 sensors across a city that are writing data every second. Now, you're talking about millions of data points per minute that need to be written into a database, and most [general-purpose] databases don't do that well.
A second thing is that when you have that data and know what's happening over time, you often want to build sophisticated control systems and reporting around the data. You want to be able to query that data really fast. If it arrives in a second, you want to be able to respond to it another second. If you get a weird data point that could put you in a critical condition, you want to see that right away. You're therefore querying the data constantly, and that data query has to be faster than normal. You're querying aggressively at that leading edge of data, and that's real-time analytics.
What are examples where that kind of speed is necessary?
Kaplan: It could be anything from self-driving cars to tracking energy -- it's anything that has to happen in real time and requires self-correcting systems. Now, we're seeing IoT sensors everywhere -- cities, healthcare, homes. One of my children plays soccer, and soccer players wear devices that capture how far they run and their heart rate in relation to how far and how hard they run throughout the course of the game. Because everything is being measured, the number of applications are exploding.
Paul Dix, who co-founded InfluxData, started building time series databases for financial firms that were trading. In high-frequency trading, where the latency between the time a trade is issued and completed [affects value], time series is super important. In a general-purpose database, he would have had to build things to make them high-performing for a specific application, so he saw an opportunity to develop something purpose-built for time series. Now, there are a lot of competitors.
What's another advantage of a database built for time series?
Kaplan: Another thing is that because you're collecting so many measurements, you don't necessarily want to keep all of them around forever. You don't need to know what happened every half-second. You can summarize that data. A sample summarization layer can show, on average, what is happening every minute. You can do a lot of that post-processing and then evict a lot of data -- get rid of it because you don't need the original source data once it's been summarized. A time series database, a platform that's been organized around that, can be super-performing.
Evan KaplanCEO, InfluxData
What does raising $81 million in new funding enable InfluxData to do that it previously couldn't?
Kaplan: I think it keeps us on our track. The capital market has really changed. It's not a grow-at-all-cost market anymore. It's now about profitable growth, and the funding puts us on an increasing path toward profitability and gives us a cash cushion. We hadn't raised new funding since 2019, so we're pretty reasonable about our growth rate. We think of this as an important round to make sure we can continue to grow at a good pace.
We recently introduced a new database engine called IOx, and that's an important key to our future around performance and compression of the data. The funding helps us continue to invest in IOx and other product development.
Are there specific plans yet for how the funds will be earmarked?
Kaplan: It's primarily oriented for product development, continuing to add to the open source and the things we contribute to evolve the platform. Of course, there's also sales and marketing, which as you grow, you have to add. There's also between 10% and 15% people growth. We want to be judicious. We want healthy growth, but not like two or three years ago when people were adding personnel.
Speaking of now versus two or three years ago, was it difficult to raise funding given venture capitalist investments in data and analytics slowed in 2022 compared with 2021 and before?
Kaplan: I think there is a long-term belief that independent of the normal cycles of technology, that this space around real-time analytics -- time series stuff -- is a long-term grower. So investors that are not looking for a quick return, but see value over a long period of time see value in the market. It made it easier for us than it might have been otherwise, but it was still not as easy as it was in 2019.
The team has been through [cycles] before. We've made mistakes. But we have a reasonably adult view of what it will take to grow this business over a long period of time as opposed to just the next 12 months.
Beyond the funding news, as InfluxData continues to evolve, what is its greatest strength and what is an area where it could use improvement?
Kaplan: I think our greatest strength is willingness to try things out, make mistakes, embrace failure and to move ahead. We've made some significant bets in the history of the company, and some have paid off and some haven't. But we're all oriented that way in technology. I also think a strength is deep understanding of use cases and problems because we have a huge open source community that tells us what they're doing and how their needs are evolving. That's the exciting part.
In terms of what we could do better, that's an endless list. Historically, we've been strong at working alone -- working directly with our customers. We've had some really important partnerships, but could be better at partnering with some other vendors that could complement what we do.
Editor's note: This Q&A has been edited for clarity and conciseness.
Eric Avidon is a senior news writer for TechTarget Editorial and is a journalist with more than 25 years of experience. He covers analytics and data management.