StarZero All shows Try the podcast clipper →

May 27, 2014 · 23m · mad

Ashish Thusoo, Qubole // Data Driven #26 // April 2014 (Hosted by FirstMark Capital)

Ashish Thusoo · 19m spoken Matt Turck · 46s spoken Jerry Galar · 26s spoken David Kim · 22s spoken
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 Data Driven NYC event, Qubole CEO Ashish Thusoo explains how cloud-native big data infrastructure transforms enterprise analytics by offering elasticity, financial flexibility, and operational agility over traditional on-premise Hadoop deployments.

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 3.7% of the talking time here. How this is scored →

Matt as informed peer 1.2 Guest teaching 5.8 Guest disagreement 0.8 Matt pushing back 0.7
05100:0010:0020:002:11–6:29 · Matt as informed peer 0/10 Comparing On-Premise Friction with Cloud Agility As this is a guest presentation monologue, host expertise and pushback are zero. Ashish educates the audience on the structural friction of on-premise Hadoop hardware procurement compared to instant cloud elasticity.6:29–9:29 · Matt as informed peer 0/10 Risk Mitigation, Cost Models, and Compute Flexibility In this monologue segment, the host does not intervene. Ashish explains cost risk mitigation and compute flexibility, detailing how decoupled storage and compute on AWS enable flexible rental models.9:29–13:48 · Matt as informed peer 0/10 Operational Management and Elastic Scalability During this monologue section, host scores remain zero. Ashish draws on his experience running Facebook-scale analytics clusters to contrast legacy capacity planning with cloud elasticity.13:48–15:51 · Matt as informed peer 0/10 Key Criteria for Evaluating Cloud Hadoop Services Ashish wraps up his talk by warning against running static long-lived Hadoop clusters in the cloud. Host involvement is absent during this final presentation segment.15:51–18:54 · Matt as informed peer 5/10 Q&A: Enterprise Cloud Readiness and TCO Inflection Points Matt Turck opens the Q&A by citing on-premise Hadoop competitors who claim Fortune 500 enterprises will never move to the cloud. Ashish respectfully reframes this by comparing enterprise cloud adoption to developing nations bypassing landlines for mobile phones.18:54–23:29 · Matt as informed peer 2/10 Q&A: Cloud Economics and Managed Services for Startups Matt facilitates audience questions from Jerry Galar and David Kim. Ashish forcefully dismisses the idea of early-stage startups running DIY Hadoop clusters, arguing it wastes scarce engineering resources.2:11–6:29 · Guest teaching 6/10 Comparing On-Premise Friction with Cloud Agility As this is a guest presentation monologue, host expertise and pushback are zero. Ashish educates the audience on the structural friction of on-premise Hadoop hardware procurement compared to instant cloud elasticity.6:29–9:29 · Guest teaching 6/10 Risk Mitigation, Cost Models, and Compute Flexibility In this monologue segment, the host does not intervene. Ashish explains cost risk mitigation and compute flexibility, detailing how decoupled storage and compute on AWS enable flexible rental models.9:29–13:48 · Guest teaching 6/10 Operational Management and Elastic Scalability During this monologue section, host scores remain zero. Ashish draws on his experience running Facebook-scale analytics clusters to contrast legacy capacity planning with cloud elasticity.13:48–15:51 · Guest teaching 6/10 Key Criteria for Evaluating Cloud Hadoop Services Ashish wraps up his talk by warning against running static long-lived Hadoop clusters in the cloud. Host involvement is absent during this final presentation segment.15:51–18:54 · Guest teaching 5/10 Q&A: Enterprise Cloud Readiness and TCO Inflection Points Matt Turck opens the Q&A by citing on-premise Hadoop competitors who claim Fortune 500 enterprises will never move to the cloud. Ashish respectfully reframes this by comparing enterprise cloud adoption to developing nations bypassing landlines for mobile phones.18:54–23:29 · Guest teaching 6/10 Q&A: Cloud Economics and Managed Services for Startups Matt facilitates audience questions from Jerry Galar and David Kim. Ashish forcefully dismisses the idea of early-stage startups running DIY Hadoop clusters, arguing it wastes scarce engineering resources.2:11–6:29 · Guest disagreement 0/10 Comparing On-Premise Friction with Cloud Agility As this is a guest presentation monologue, host expertise and pushback are zero. Ashish educates the audience on the structural friction of on-premise Hadoop hardware procurement compared to instant cloud elasticity.6:29–9:29 · Guest disagreement 0/10 Risk Mitigation, Cost Models, and Compute Flexibility In this monologue segment, the host does not intervene. Ashish explains cost risk mitigation and compute flexibility, detailing how decoupled storage and compute on AWS enable flexible rental models.9:29–13:48 · Guest disagreement 0/10 Operational Management and Elastic Scalability During this monologue section, host scores remain zero. Ashish draws on his experience running Facebook-scale analytics clusters to contrast legacy capacity planning with cloud elasticity.13:48–15:51 · Guest disagreement 0/10 Key Criteria for Evaluating Cloud Hadoop Services Ashish wraps up his talk by warning against running static long-lived Hadoop clusters in the cloud. Host involvement is absent during this final presentation segment.15:51–18:54 · Guest disagreement 2/10 Q&A: Enterprise Cloud Readiness and TCO Inflection Points Matt Turck opens the Q&A by citing on-premise Hadoop competitors who claim Fortune 500 enterprises will never move to the cloud. Ashish respectfully reframes this by comparing enterprise cloud adoption to developing nations bypassing landlines for mobile phones.18:54–23:29 · Guest disagreement 3/10 Q&A: Cloud Economics and Managed Services for Startups Matt facilitates audience questions from Jerry Galar and David Kim. Ashish forcefully dismisses the idea of early-stage startups running DIY Hadoop clusters, arguing it wastes scarce engineering resources.2:11–6:29 · Matt pushing back 0/10 Comparing On-Premise Friction with Cloud Agility As this is a guest presentation monologue, host expertise and pushback are zero. Ashish educates the audience on the structural friction of on-premise Hadoop hardware procurement compared to instant cloud elasticity.6:29–9:29 · Matt pushing back 0/10 Risk Mitigation, Cost Models, and Compute Flexibility In this monologue segment, the host does not intervene. Ashish explains cost risk mitigation and compute flexibility, detailing how decoupled storage and compute on AWS enable flexible rental models.9:29–13:48 · Matt pushing back 0/10 Operational Management and Elastic Scalability During this monologue section, host scores remain zero. Ashish draws on his experience running Facebook-scale analytics clusters to contrast legacy capacity planning with cloud elasticity.13:48–15:51 · Matt pushing back 0/10 Key Criteria for Evaluating Cloud Hadoop Services Ashish wraps up his talk by warning against running static long-lived Hadoop clusters in the cloud. Host involvement is absent during this final presentation segment.15:51–18:54 · Matt pushing back 4/10 Q&A: Enterprise Cloud Readiness and TCO Inflection Points Matt Turck opens the Q&A by citing on-premise Hadoop competitors who claim Fortune 500 enterprises will never move to the cloud. Ashish respectfully reframes this by comparing enterprise cloud adoption to developing nations bypassing landlines for mobile phones.18:54–23:29 · Matt pushing back 0/10 Q&A: Cloud Economics and Managed Services for Startups Matt facilitates audience questions from Jerry Galar and David Kim. Ashish forcefully dismisses the idea of early-stage startups running DIY Hadoop clusters, arguing it wastes scarce engineering resources.

speaking balance: gold is Matt, purple is the guest (3 minute bins)

0:00 · Matt 0.2% · guest 99.8%0:00 · Matt 0.2% · guest 99.8%3:00 · Matt 0% · guest 100%3:00 · Matt 0% · guest 100%6:00 · Matt 0% · guest 100%6:00 · Matt 0% · guest 100%9:00 · Matt 0% · guest 100%9:00 · Matt 0% · guest 100%12:00 · Matt 0% · guest 100%12:00 · Matt 0% · guest 100%15:00 · Matt 18.7% · guest 81.3%15:00 · Matt 18.7% · guest 81.3%18:00 · Matt 10.4% · guest 89.6%18:00 · Matt 10.4% · guest 89.6%21:00 · Matt 0.5% · guest 99.5%21:00 · Matt 0.5% · guest 99.5%
Sharpest disagreement ▶ 21:28 Strong rejection of DIY cloud Hadoop

Ashish forcefully rejects the premise that a small startup should manage its own cloud Hadoop, arguing it diverts core engineering focus away from building value.

 Hardest push from Matt ▶ 15:54 Host pushes on enterprise skepticism

Matt Turck challenges Qubole's market outlook by bringing up industry claims that Fortune 500 companies will never adopt cloud-hosted Hadoop solutions.

 Biggest teaching moment ▶ 18:41 Cloud cost inflection dynamics

Ashish educates the room on cloud TCO dynamics, detailing recent 60-70% cloud price drops and explaining how scale pushes the on-premise cost inflection point outward.

 Matt holds his own ▶ 15:54 Host cites competing market perspectives

Matt Turck demonstrates deep industry knowledge by citing specific arguments from on-premise Hadoop vendors regarding enterprise data security and cloud hesitation.
the scores for every segment, with the reasoning behind each
ChapterTopicMatt as informed peerGuest teachingGuest disagreementMatt pushing backWhy
Comparing On-Premise Friction with Cloud Agility 0600 As this is a guest presentation monologue, host expertise and pushback are zero. Ashish educates the audience on the structural friction of on-premise Hadoop hardware procurement compared to instant cloud elasticity.
Risk Mitigation, Cost Models, and Compute Flexibility 0600 In this monologue segment, the host does not intervene. Ashish explains cost risk mitigation and compute flexibility, detailing how decoupled storage and compute on AWS enable flexible rental models.
Operational Management and Elastic Scalability 0600 During this monologue section, host scores remain zero. Ashish draws on his experience running Facebook-scale analytics clusters to contrast legacy capacity planning with cloud elasticity.
Key Criteria for Evaluating Cloud Hadoop Services 0600 Ashish wraps up his talk by warning against running static long-lived Hadoop clusters in the cloud. Host involvement is absent during this final presentation segment.
Q&A: Enterprise Cloud Readiness and TCO Inflection Points 5524 Matt Turck opens the Q&A by citing on-premise Hadoop competitors who claim Fortune 500 enterprises will never move to the cloud. Ashish respectfully reframes this by comparing enterprise cloud adoption to developing nations bypassing landlines for mobile phones.
Q&A: Cloud Economics and Managed Services for Startups 2630 Matt facilitates audience questions from Jerry Galar and David Kim. Ashish forcefully dismisses the idea of early-stage startups running DIY Hadoop clusters, arguing it wastes scarce engineering resources.

Statements from this episode (6)

Disclosure
Qubole runs cloud Hadoop clusters for Pinterest, Quora, and MediaMath
“As part of QBOL, we run, ah, very large Hadoop clusters on the cloud for, you know, companies like Pinterest Quora, MediaMath, DataZoo, and so on and so forth.”
Ashish Thusoo May 27, 2014 ▶ 0:47
Assertion Not checkable as stated
Facebook data infrastructure capacity planning consistently lagged behind demand
“Having, you know, run big clusters at Facebook, I remember conversations with the procurement team where it was like a big capacity planning exercise trying to predict the future, and yes, I know predictive analytics works to predict the future really well, bu…”
Ashish Thusoo May 27, 2014 ▶ 11:00
Insight
Static Hadoop clusters in the cloud defeat the purpose of elasticity
“They just run long-running Hadoop clusters, which completely defeat the purpose of You know, how, how you can leverage the cloud to be dynamically adaptable to your workloads and things like that.”
Ashish Thusoo May 27, 2014 ▶ 15:15
Insight
Public cloud does not make financial sense at Facebook scale
“So that's an excellent question, and, ah, there is definitely an, there, ah, you know, there is an inflection point where people, ah, think that, ah, the total cost of ownership, ah, you know, once you're, ah, so suppose you're a Facebook skill, right, of cour…”
Ashish Thusoo May 27, 2014 ▶ 18:42
Assertion Partly supported
Cloud storage and compute prices recently dropped 70% and 60%
“The costs, the storage costs in the cloud, for example, S-three and so on and so forth, were cut to one third, like a 70% drop. Same for compute, you know, 60% drop.”
Ashish Thusoo May 27, 2014 ▶ 19:15
Assertion Not checkable as stated
Unmodified on-premise Hadoop distributions fail in the cloud beyond 10 nodes
“If you just take a normal Hadoop distro and try to run it in the cloud, the chances are at 10 nodes it'll work fine as you start growing and, you know, as you start growing and growing and growing further. Things will start breaking because, you know, compute …”
Ashish Thusoo May 27, 2014 ▶ 22:34
Made with StarZero

Turn any episode into a week of clips.

This entire site, over 400 conversations transcribed, diarized, checked and made playable, runs on the StarZero media pipeline. Drop in your own episode and the podcast clipper finds the moments worth sharing, cuts them, captions them, and reframes them for every feed.

Try the podcast clipper Learn more