Everything Ashish Thusoo said on any show that made the record, most notable first. Each card names its show and opens the statement there.
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…”
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 …”
Thusoo: Centralized enterprise data teams usually become operational bottlenecks
“More often than not, this, ah, data team eventually becomes a bottleneck, ah, and, ah, you know, it sort of delays, ah, delays the, ah, availability of data and what people can do with that data.”
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.”
Thusoo: On-premise big data deployments take 6 to 18 months
“If you are, you know, going with a traditional approach of, you know, standing this up in your data centers, you know, ah, I think some of the estimates are it takes between six to 18 months to get to production.”
Thusoo: Cloud infrastructure pricing decreased between 2012 and 2016
“Actually the cloud pricing has gone down over the last three, four years.”
Thusoo: Multi-cloud enterprise adoption remains rare in 2016
“We haven't seen, we have started to see in the press mentions of a company using two clouds and stuff like that, but it's not very wholesale right now.”
Thusoo: Cloud vendor lock-in concerns will diminish as cloud-agnostic software grows
“I think some of those vendor lock-in issues, they'll start to go away.”
Thusoo: Qubole clusters regularly auto-scale 36x for average customers
“QBOL, for example, in QBOL, you know, we regularly have, ah, you know, with our autoscaling algorithms, we are, we regularly scale up clusters to 36 times the size of the minimum clusters on an average customer”
Qubole processes around 500 petabytes of data monthly
“You know, for example, we process around 500 petabytes of data every month across workloads ranging from you know, mundane ETL workloads to more complex machine learning workloads to ad hoc analysis and so on and so forth.”
Capital One publicly announced plans to close data centers at AWS re:Invent
“I think last year in reInvent, right, Capital One was up in the state saying that, you know, I'm going to, we're going to shut down our data centers, move things to the cloud.”
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.”
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…”
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 led Facebook's data infrastructure team before founding Qubole
“Before I started QBOL, I was at Facebook, and I built out, I ran the team that built out the infrastructure there.”
the MAD Podcast 1