Dec 17, 2015 · 35m · mad
A Fireside Chat with MapR CTO M.C. Srivas (Data Driven NYC / FirstMark)
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
At a FirstMark Data Driven NYC event, MapR co-founder and CTO M.C. Srivas joins Matt Turck to discuss the founding, architectural innovations, and real-world enterprise applications of MapR's big data platform. He shares insights on big data scale, open-source strategy, and major implementations like India's 1.3-billion-person Aadhaar biometric project.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Matt holds 10.1% of the talking time here. How this is scored →
Matt as informed peer 2.4 Guest teaching 4.1 Guest disagreement 1.4 Matt pushing back 1.3
speaking balance: gold is Matt, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 22:49 Exposing open source hypocrisies
Srivas forcefully rejects the open-source framing promoted by competitors like Cloudera and Hortonworks, calling their claims disingenuous and exposing how vendors privately control open-source projects.
Hardest push from Matt ▶ 21:29 Challenging MapR's competitive positioningMatt cites Cloudera founder Mike Olson and Hortonworks' open-source strategy to directly challenge Srivas on how MapR differentiates in a ruthlessly competitive market.
Biggest teaching moment ▶ 8:29 Explaining HDFS architectural limitsSrivas clearly educates the room by detailing exact numerical limitations of HDFS clusters regarding file counts and latency, illustrating why native architecture was required for high-scale applications.
Matt holds his own ▶ 21:29 Framing the open-source competitive landscapeMatt demonstrates high domain expertise by invoking specific rival executives and distinct open-source philosophical stances to force Srivas into defending MapR's architecture.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Matt as informed peer | Guest teaching | Guest disagreement | Matt pushing back | Why |
|---|---|---|---|---|---|---|
| MapR Founding Story and Early Startup Challenges | 2 | 3 | 1 | 1 | Matt sets up the founding story by noting MapR's funding and headcount, then asks standard questions about hiring and early sales. Srivas explains early struggles, like hiring in a sparse room without chairs and learning that potential customers lie during pre-product interviews. | |
| MapRFS vs HDFS and High-Scale Email Infrastructure | 3 | 5 | 1 | 1 | Matt asks a direct technical question about how MapRFS differs from HDFS. Srivas educates the room by detailing HDFS's architectural limits with small files and scale, contrasting it with MapRFS handling massive email traffic and multi-temperature storage. | |
| MapRDB and India's Aadhaar Biometric Identification System | 1 | 4 | 0 | 0 | Matt prompts Srivas to discuss MapRDB, leading to an extended story from Srivas about NoSQL history and India's Aadhaar biometric system. Srivas dominates the narrative detailing how MapRDB powers identity verification for over 900 million citizens. | |
| MapR Streams and Real-Time IoT Data Processing | 1 | 4 | 1 | 0 | Matt introduces MapR Streams, and Srivas leads an engaging discussion on real-time streaming for IoT. Srivas quizzes the audience and explains the massive data footprint of self-driving cars and defense systems. | |
| Open Source Strategy and Competitive Landscape | 6 | 6 | 5 | 5 | Matt pushes Srivas on the tension between open source and proprietary models and highlights competitors like Cloudera and Hortonworks. Srivas aggressively reframes the debate, accusing competitors of being disingenuous and exposing single-vendor control over open source projects. | |
| Mainstream Adoption and Future of Big Data | 3 | 3 | 1 | 1 | Matt asks if Hadoop has crossed the chasm into mainstream enterprise adoption. Srivas explains that flexible schema requirements have made big data a standard budget item for major companies before transitioning to audience questions. | |
| Q&A: Google Technical Leadership and Native File Systems | 1 | 4 | 1 | 1 | Audience members ask about Google's technical lead and MapR's native file system capabilities outside Hadoop. Srivas provides technical perspective on Google's historic advantage and confirms MapRFS usage via POSIX and NFS interfaces. |
Statements from this episode (20)
Turck: MapR raised a $110 million funding round from Google Capital
“Most recently it was a hundred and ten million round from Google capital.”
Srivas: Early founders must intentionally avoid hiring people like themselves
“And you have to try to hire people who are not like you. I mean, basically, that's the first advice. Because you tend to be attracted to people who are like you, and in the beginning you really want a big variety of people, and not from the same DNA.”
Srivas interviewed 50 Hadoop-using companies before founding MapR
“You know, well, before we started Mapper, I spoke to, like, about 40 or 50 people who were using Hadoop. 50 companies.”
Srivas: First funding rounds build product; later rounds fund pure sales
“Because the way this venture money flows is that you get the first bunch of money for building the product, and not again. After that, you get the money for selling it.”
Srivas: MapR powers a global email provider with 1.5 billion accounts
“We are now, from a file system perspective, we are powering one of the largest email providers in the world. Ok, ah, when I say largest email providers, everybody of you probably is on it, ah, one and a half billion email accounts.”
Srivas: A single Apache HDFS cluster handles roughly 100 million files
“HDFS, a single cluster, can do about a hundred million files.”
Srivas: Aadhaar will take 2.5 years to achieve full enrollment in India
“At the rate of one and a half million a day, they're gonna go for another two and a half years before everybody's on.”
Srivas: India's Aadhaar biometric database runs on MapR across four datacenters
“Now, this is running on MapRDB in four data centers worldwide, in across India.”
Srivas: MapR provided its database software to India's Aadhaar project for free
“We gave it for free, by the way, right, because we thought it was very important.”
Srivas: A single self-driving car produces one terabyte of data hourly
“So it's a single self-driving car produces one terabyte per hour.”
Srivas: Next scale of big data requires thousands of edge processing clusters
“I think the real next scale in, in Stefan's graph is you need thousands of clusters that do edge processing everywhere.”
Srivas: Cloudera and Hortonworks are disingenuous about their open source control
“I think that's a bit disingenuous on Hortonworks and Cloudrass part to say that Because they would like to, so the, here's the dirty little truth about open source software, open source vendors, right? I love open source as long as it's my open source and not …”
Srivas: Open source does not guarantee high software quality
“It's not clear that open source equals good quality yet.”
Srivas: MapR has reached 1,000 customers in four years
“We have about a thousand customers now, right? In four years.”
Srivas: Almost every enterprise now has a Hadoop budget item
“Every company now has a Hadoop budget item. Almost every company.”
Srivas: The era of rigid database schemas is over
“The old schema, rigid schema is gone. I mean, if you're still waiting, thinking that it's gonna come back, it's not. It's over.”
Srivas: India's Aadhaar biometric system has registered 930 million people
“There's nine hundred and thirty million people online right now.”