SlideShare a Scribd company logo
ScyllaDB
Beyond Cassandra
Dor Laor
C10K
C10K
D10M/
$10M
D10M/
$10M
Scylla Summit 2019 Keynote - Dor Laor - Beyond Cassandra
Scylla Summit 2019 Keynote - Dor Laor - Beyond Cassandra
Hello NoSQL...
Scylla Summit 2019 Keynote - Dor Laor - Beyond Cassandra
Scale on..
Hello DBaaS
Node
Sprawl
Compact 45TB/node at 8GB/s SLA for 3 workloads @ 300k OPS
Only Scylla Can
Let’s Test
Our Limits
Scylla and Packet:
Reaching new heights
Glauber Costa, VP Field Engineering @ Scylla
James Malachowski, Solutions Architect @ Packet
An IoT application
Total amount of data points
526 billion
temperature readings
1,000,000 sensors, representing homes in an area
365 days (1 year storage requirement) 1 reading per minute
Data Model
CREATE TABLE readings (
sensor_id int,
date date,
time time,
temperature float,
PRIMARY KEY ((sensor_id, date), time))
What kind of queries can we reasonably support?
■ SELECT * from readings where sensor_id = ? and date = ?;
■ SELECT * from readings where sensor_id = ? and date = ? and time > ?;
Analytics over the entire data?
How long would it take
at normal speeds?
We need more if
analytics are a part of the
pipeline
That means we need Scylla
We need a good application
And we need hardware
200,000 points/second
730 hours (30 days)
1 million points/second
146 hours (almost a week)
Why climb Mount Everest?
Because it’s there.
George Leigh Mallory
What kind of performance are we after?
Analytics Application
SELECT count(temperature) as totalRows,
sum(temperature) as sumTemperature,
min(temperature) as minTemperature,
max(temperature) as maxTemperature
FROM readings where sensor_id = ? and date = ?`
Application
(Example) Total amount of data to scan: 1.44 billion points/day
Coordinator
Worker
(loader machine)
ScyllaDB cluster
Worker
(loader machine)
Worker
(loader machine)
Set time frame,
compute average,
min, max of
all sensors
Packet is a global bare metal cloud that is built for Enterprises,
and loved by Developers.
TLDR: We sell servers to millennials.
Core Edge
Singapore
Sydney
Seattle
Sunnyvale
Dallas
Amsterdam
Los Angeles
New York
Tokyo
Ashburn
Atlanta
Toronto
Frankfurt
Chicago
Hong Kong
Marseille
Dublin
22+ Locations + Custom Deployments
The Database
(83 Machines)
The Workers
(24 Machines)
Deployed via
Packet’s API using
Ansible playbooks.
83 Database Servers
2800 Physical Cores
Intel Cascale Lake
314 TB of NVMe
34 TB of RAM
Results
Scanning 3 months of data
Finished Scanning. Succeeded 132,480,000,000 rows. Failed 0 rows. RPC Failures: 0. Took 110,892.91 ms
Processed 1,194,666,083 rows/s
Absolute min: 19.71, date 2019-08-27, sensorID 473869
Absolute max: 135.21, date 2019-08-27, sensorID 473869
Results
Scanning 3 months of data
Finished Scanning. Succeeded 132,480,000,000 rows. Failed 0 rows. RPC Failures: 0. Took 110,892.91 ms
Processed 1,194,666,083 rows/s
Absolute min: 19.71, date 2019-08-27, sensorID 473869
Absolute max: 135.21, date 2019-08-27, sensorID 473869
Results
Scanning 3 months of data
Finished Scanning. Succeeded 132,480,000,000 rows. Failed 0 rows. RPC Failures: 0. Took 110,892.91 ms
Processed 1,194,666,083 rows/s
Absolute min: 19.71, date 2019-08-27, sensorID 473869
Absolute max: 135.21, date 2019-08-27, sensorID 473869
Results
Scanning the entire dataset
Finished Scanning. Succeeded 525,599,474,400 rows. Failed 0 rows. RPC Failures: 0. Took 542,191.31 ms
Processed 969,398,554 rows/s
Absolute min: 68.00, date 2019-05-28, sensorID 82114
Absolute max: 79.99, date 2019-03-19, sensorID 152594
Results
Scanning the entire dataset
Finished Scanning. Succeeded 525,599,474,400 rows. Failed 0 rows. RPC Failures: 0. Took 542,191.31 ms
Processed 969,398,554 rows/s
Absolute min: 68.00, date 2019-05-28, sensorID 82114
Absolute max: 79.99, date 2019-03-19, sensorID 152594
Results
Scanning the entire dataset
Finished Scanning. Succeeded 525,599,474,400 rows. Failed 0 rows. RPC Failures: 0. Took 542,191.31 ms
Processed 969,398,554 rows/s
Absolute min: 68.00, date 2019-05-28, sensorID 82114
Absolute max: 79.99, date 2019-03-19, sensorID 152594
We summited
Mount Everest
(and survived)
We can efficiently process
over 1.2 billion points
per second
(we’ll process whatever you need, too!)
Bare metal + Scylla are
like Peanut Butter & Jelly.
+
Packet & ScyllaDB together, provide a scalable
& low latency database, deployed on a global
bare metal cloud!
“All Companies are
Software Companies”
Satya Nadella
“All Companies are Software”
“All Companies are
Software Companies”
Jay Kreps
“All Companies are Software”
“All Companies are Software”
“All Companies are Software”
Make it EASY for software companies
Make it EASY for software companies
■ Scylla Cloud
Make it EASY for software companies
■ Scylla Cloud
■ Project Alternator
Scylla gains DynamoDB compatible API
■ Scylla Cloud
■ Project Alternator
Scylla gains DynamoDB compatible API
Make it EASY for software companies
120 Kops/sec is pretty intensive -
in DynamoDB provisioned pricing,
it would cost $85 per hour.
VMs (EC2, on-demand pricing) for
running Scylla - $7.5 per hour.
Year reservation cheaper for both.
Make it EASY for software companies
Scylla Cloud
Workload
prioritization
Scylla
Manager
BackupAlternator
Make it EASY for software companies
LWT
Scylla Cloud
Workload
prioritization
Scylla
Manager
BackupAlternator
Make it EASY for software companies
LWT
Scylla Cloud
Workload
prioritization
Scylla
Manager
BackupAlternator
CDC
Make it EASY for software companies
UDF/
TransformationsLWT
Scylla Cloud
Workload
prioritization
Scylla
Manager
BackupAlternator
CDC
Make it EASY for software companies
Scylla & ScyllaDB
■ Major Round of Funding
■ 2x Engineers
■ 2x Customers
■ $ you spend goes back to you
Func6 Customer story
■ Required by a new customer
■ Although code ready, it wasn’t yet tested
■ We had to double down to release it
Func6 Customer story
Greg, account
executive
Eyal coordinated
the requirement
Shlomi, assigned
people
Luis, tested the
OSS version
Pekka built +
back ported to
enterprise
Calle fixed bugs
Gleb reviewed
code
Kamil reviewed
& tested
Michal designed,
Fixed, built Scylla
Manager
Amnon adjusted
the monitor
Amos tested
Roy tested Bentsi tested
Julia tested
“All Companies are Software Companies”
“Software is Eating the World”
“All Companies are Software”
Scylla is a damn good
Choice for *your* Software
Stay in touch
@DorLaor
Dor Laor
dor@scylladb.com
Where are we going
Backup - point in time, auto-pre-upgrade
Tablets
k8s/manager
Release Schedule
2nd Quarter 2020 3rd Quarter 2020
Apr May Jun Jul Aug Sep
4th Quarter 2019
Oct Nov Dec
1st Quarter 2020
Jan Feb Mar
Scylla
Enterprise
2019.1.3
Scylla Manager 2.0
Scylla 3.2
Scylla Enterprise
2019.1.4
Scylla
Enterprise
2020.1
Scylla 4.0
Scylla Manager
2.1
Scylla Manager
3.0

More Related Content

PPTX
How to be Successful with Scylla
PPTX
iFood on Delivering 100 Million Events a Month to Restaurants with Scylla
PPTX
Sizing Your Scylla Cluster
PPTX
How Opera Syncs Tens of Millions of Browsers and Sleeps Well at Night
PPTX
FireEye & Scylla: Intel Threat Analysis Using a Graph Database
PDF
Scylla: 1 Million CQL operations per second per server
PPTX
How Workload Prioritization Reduces Your Datacenter Footprint
PDF
ScyllaDB @ Apache BigData, may 2016
How to be Successful with Scylla
iFood on Delivering 100 Million Events a Month to Restaurants with Scylla
Sizing Your Scylla Cluster
How Opera Syncs Tens of Millions of Browsers and Sleeps Well at Night
FireEye & Scylla: Intel Threat Analysis Using a Graph Database
Scylla: 1 Million CQL operations per second per server
How Workload Prioritization Reduces Your Datacenter Footprint
ScyllaDB @ Apache BigData, may 2016

What's hot (20)

PDF
Lookout on Scaling Security to 100 Million Devices
PDF
Looking towards an official cassandra sidecar netflix
PDF
Scylla Summit 2022: Building Zeotap's Privacy Compliant Customer Data Platfor...
PDF
Scylla Summit 2016: Graph Processing with Titan and Scylla
PPTX
Writing Applications for Scylla
PPTX
Scylla Summit 2022: Scylla 5.0 New Features, Part 1
PPTX
Scylla’s Journey Towards Being an Elastic Cloud Native Database
PDF
Scylla Summit 2022: Operating at Monstrous Scales: Benchmarking Petabyte Work...
PPTX
Scylla Summit 2018: Meshify - A Case Study, or Petshop Seamonsters
PDF
Scylla Summit 2016: Analytics Show Time - Spark and Presto Powered by Scylla
PPTX
Seastar Summit 2019 Keynote
PPTX
Scylla Summit 2018: Scylla Feature Talks - Scylla Streaming and Repair Updates
PPTX
Scylla Summit 2018: OLAP or OLTP? Why Not Both?
PPTX
Lightweight Transactions at Lightning Speed
PDF
How to Monitor and Size Workloads on AWS i3 instances
PDF
Renegotiating the boundary between database latency and consistency
PPTX
Scylla Summit 2019 Keynote - Avi Kivity
PPTX
Connecting kafka message systems with scylla
PPTX
SAS Institute on Changing All Four Tires While Driving an AdTech Engine at Fu...
PDF
Seastar Summit 2019 vectorized.io
Lookout on Scaling Security to 100 Million Devices
Looking towards an official cassandra sidecar netflix
Scylla Summit 2022: Building Zeotap's Privacy Compliant Customer Data Platfor...
Scylla Summit 2016: Graph Processing with Titan and Scylla
Writing Applications for Scylla
Scylla Summit 2022: Scylla 5.0 New Features, Part 1
Scylla’s Journey Towards Being an Elastic Cloud Native Database
Scylla Summit 2022: Operating at Monstrous Scales: Benchmarking Petabyte Work...
Scylla Summit 2018: Meshify - A Case Study, or Petshop Seamonsters
Scylla Summit 2016: Analytics Show Time - Spark and Presto Powered by Scylla
Seastar Summit 2019 Keynote
Scylla Summit 2018: Scylla Feature Talks - Scylla Streaming and Repair Updates
Scylla Summit 2018: OLAP or OLTP? Why Not Both?
Lightweight Transactions at Lightning Speed
How to Monitor and Size Workloads on AWS i3 instances
Renegotiating the boundary between database latency and consistency
Scylla Summit 2019 Keynote - Avi Kivity
Connecting kafka message systems with scylla
SAS Institute on Changing All Four Tires While Driving an AdTech Engine at Fu...
Seastar Summit 2019 vectorized.io
Ad

Similar to Scylla Summit 2019 Keynote - Dor Laor - Beyond Cassandra (20)

PDF
The True Cost of NoSQL DBaaS Options
PPTX
A Deep Dive into ScyllaDB's Architecture
PDF
Scylla Summit 2022: How ScyllaDB Powers This Next Tech Cycle
PPTX
Scylla Summit 2018: Joining Billions of Rows in Seconds with One Database Ins...
PDF
Use ScyllaDB Alternator to Use Amazon DynamoDB API, Everywhere, Better, More ...
PDF
ScyllaDB Virtual Workshop
PDF
Using ScyllaDB for Real-Time Read-Heavy Workloads.pdf
PDF
Using ScyllaDB for Real-Time Write-Heavy Workloads
PDF
Dissecting Real-World Database Performance Dilemmas
PDF
Introducing Scylla Cloud
PDF
Numberly on Joining Billions of Rows in Seconds: Replacing MongoDB and Hive w...
PDF
New Ways to Reduce Database Costs with ScyllaDB
PDF
AdGear Use Case with Scylla - 1M Queries Per Second with Single-Digit Millise...
PPTX
Piranha vs. mammoth predator appliances that chew up big data
PDF
To Serverless and Beyond
PPTX
4 use cases for C* to Scylla
PDF
Dissecting Real-World Database Performance Dilemmas
PDF
Inside Freshworks' Migration from Cassandra to ScyllaDB by Premkumar Patturaj
PPTX
Augury: Real-Time Insights for the Industrial IoT
PDF
How Development Teams Cut Costs with ScyllaDB.pdf
The True Cost of NoSQL DBaaS Options
A Deep Dive into ScyllaDB's Architecture
Scylla Summit 2022: How ScyllaDB Powers This Next Tech Cycle
Scylla Summit 2018: Joining Billions of Rows in Seconds with One Database Ins...
Use ScyllaDB Alternator to Use Amazon DynamoDB API, Everywhere, Better, More ...
ScyllaDB Virtual Workshop
Using ScyllaDB for Real-Time Read-Heavy Workloads.pdf
Using ScyllaDB for Real-Time Write-Heavy Workloads
Dissecting Real-World Database Performance Dilemmas
Introducing Scylla Cloud
Numberly on Joining Billions of Rows in Seconds: Replacing MongoDB and Hive w...
New Ways to Reduce Database Costs with ScyllaDB
AdGear Use Case with Scylla - 1M Queries Per Second with Single-Digit Millise...
Piranha vs. mammoth predator appliances that chew up big data
To Serverless and Beyond
4 use cases for C* to Scylla
Dissecting Real-World Database Performance Dilemmas
Inside Freshworks' Migration from Cassandra to ScyllaDB by Premkumar Patturaj
Augury: Real-Time Insights for the Industrial IoT
How Development Teams Cut Costs with ScyllaDB.pdf
Ad

More from ScyllaDB (20)

PDF
Understanding The True Cost of DynamoDB Webinar
PDF
Database Benchmarking for Performance Masterclass: Session 2 - Data Modeling ...
PDF
Database Benchmarking for Performance Masterclass: Session 1 - Benchmarking F...
PDF
Designing Low-Latency Systems with Rust and ScyllaDB: An Architectural Deep Dive
PDF
Powering a Billion Dreams: Scaling Meesho’s E-commerce Revolution with Scylla...
PDF
Leading a High-Stakes Database Migration
PDF
Achieving Extreme Scale with ScyllaDB: Tips & Tradeoffs
PDF
Securely Serving Millions of Boot Artifacts a Day by João Pedro Lima & Matt ...
PDF
How Agoda Scaled 50x Throughput with ScyllaDB by Worakarn Isaratham
PDF
How Yieldmo Cut Database Costs and Cloud Dependencies Fast by Todd Coleman
PDF
ScyllaDB: 10 Years and Beyond by Dor Laor
PDF
Reduce Your Cloud Spend with ScyllaDB by Tzach Livyatan
PDF
Migrating 50TB Data From a Home-Grown Database to ScyllaDB, Fast by Terence Liu
PDF
Vector Search with ScyllaDB by Szymon Wasik
PDF
Workload Prioritization: How to Balance Multiple Workloads in a Cluster by Fe...
PDF
Two Leading Approaches to Data Virtualization, and Which Scales Better? by Da...
PDF
Scaling a Beast: Lessons from 400x Growth in a High-Stakes Financial System b...
PDF
Object Storage in ScyllaDB by Ran Regev, ScyllaDB
PDF
Lessons Learned from Building a Serverless Notifications System by Srushith R...
PDF
A Dist Sys Programmer's Journey into AI by Piotr Sarna
Understanding The True Cost of DynamoDB Webinar
Database Benchmarking for Performance Masterclass: Session 2 - Data Modeling ...
Database Benchmarking for Performance Masterclass: Session 1 - Benchmarking F...
Designing Low-Latency Systems with Rust and ScyllaDB: An Architectural Deep Dive
Powering a Billion Dreams: Scaling Meesho’s E-commerce Revolution with Scylla...
Leading a High-Stakes Database Migration
Achieving Extreme Scale with ScyllaDB: Tips & Tradeoffs
Securely Serving Millions of Boot Artifacts a Day by João Pedro Lima & Matt ...
How Agoda Scaled 50x Throughput with ScyllaDB by Worakarn Isaratham
How Yieldmo Cut Database Costs and Cloud Dependencies Fast by Todd Coleman
ScyllaDB: 10 Years and Beyond by Dor Laor
Reduce Your Cloud Spend with ScyllaDB by Tzach Livyatan
Migrating 50TB Data From a Home-Grown Database to ScyllaDB, Fast by Terence Liu
Vector Search with ScyllaDB by Szymon Wasik
Workload Prioritization: How to Balance Multiple Workloads in a Cluster by Fe...
Two Leading Approaches to Data Virtualization, and Which Scales Better? by Da...
Scaling a Beast: Lessons from 400x Growth in a High-Stakes Financial System b...
Object Storage in ScyllaDB by Ran Regev, ScyllaDB
Lessons Learned from Building a Serverless Notifications System by Srushith R...
A Dist Sys Programmer's Journey into AI by Piotr Sarna

Recently uploaded (20)

PDF
REPORT: Heating appliances market in Poland 2024
PDF
Cloud-Migration-Best-Practices-A-Practical-Guide-to-AWS-Azure-and-Google-Clou...
PDF
solutions_manual_-_materials___processing_in_manufacturing__demargo_.pdf
PDF
Revolutionize Operations with Intelligent IoT Monitoring and Control
PDF
GamePlan Trading System Review: Professional Trader's Honest Take
PDF
Dell Pro 14 Plus: Be better prepared for what’s coming
PPTX
Telecom Fraud Prevention Guide | Hyperlink InfoSystem
PDF
How-Cloud-Computing-Impacts-Businesses-in-2025-and-Beyond.pdf
PDF
Doc9.....................................
PDF
Software Development Methodologies in 2025
PDF
CIFDAQ's Market Wrap: Ethereum Leads, Bitcoin Lags, Institutions Shift
PDF
Building High-Performance Oracle Teams: Strategic Staffing for Database Manag...
PDF
NewMind AI Weekly Chronicles - July'25 - Week IV
PPTX
CroxyProxy Instagram Access id login.pptx
PDF
NewMind AI Weekly Chronicles - August'25 Week I
PPTX
breach-and-attack-simulation-cybersecurity-india-chennai-defenderrabbit-2025....
PDF
Event Presentation Google Cloud Next Extended 2025
PPTX
ABU RAUP TUGAS TIK kelas 8 hjhgjhgg.pptx
PDF
Google’s NotebookLM Unveils Video Overviews
PDF
How Onsite IT Support Drives Business Efficiency, Security, and Growth.pdf
REPORT: Heating appliances market in Poland 2024
Cloud-Migration-Best-Practices-A-Practical-Guide-to-AWS-Azure-and-Google-Clou...
solutions_manual_-_materials___processing_in_manufacturing__demargo_.pdf
Revolutionize Operations with Intelligent IoT Monitoring and Control
GamePlan Trading System Review: Professional Trader's Honest Take
Dell Pro 14 Plus: Be better prepared for what’s coming
Telecom Fraud Prevention Guide | Hyperlink InfoSystem
How-Cloud-Computing-Impacts-Businesses-in-2025-and-Beyond.pdf
Doc9.....................................
Software Development Methodologies in 2025
CIFDAQ's Market Wrap: Ethereum Leads, Bitcoin Lags, Institutions Shift
Building High-Performance Oracle Teams: Strategic Staffing for Database Manag...
NewMind AI Weekly Chronicles - July'25 - Week IV
CroxyProxy Instagram Access id login.pptx
NewMind AI Weekly Chronicles - August'25 Week I
breach-and-attack-simulation-cybersecurity-india-chennai-defenderrabbit-2025....
Event Presentation Google Cloud Next Extended 2025
ABU RAUP TUGAS TIK kelas 8 hjhgjhgg.pptx
Google’s NotebookLM Unveils Video Overviews
How Onsite IT Support Drives Business Efficiency, Security, and Growth.pdf

Scylla Summit 2019 Keynote - Dor Laor - Beyond Cassandra

  • 13. Compact 45TB/node at 8GB/s SLA for 3 workloads @ 300k OPS Only Scylla Can
  • 15. Scylla and Packet: Reaching new heights Glauber Costa, VP Field Engineering @ Scylla James Malachowski, Solutions Architect @ Packet
  • 16. An IoT application Total amount of data points 526 billion temperature readings 1,000,000 sensors, representing homes in an area 365 days (1 year storage requirement) 1 reading per minute
  • 17. Data Model CREATE TABLE readings ( sensor_id int, date date, time time, temperature float, PRIMARY KEY ((sensor_id, date), time)) What kind of queries can we reasonably support? ■ SELECT * from readings where sensor_id = ? and date = ?; ■ SELECT * from readings where sensor_id = ? and date = ? and time > ?;
  • 18. Analytics over the entire data? How long would it take at normal speeds? We need more if analytics are a part of the pipeline That means we need Scylla We need a good application And we need hardware 200,000 points/second 730 hours (30 days) 1 million points/second 146 hours (almost a week)
  • 19. Why climb Mount Everest? Because it’s there. George Leigh Mallory What kind of performance are we after?
  • 20. Analytics Application SELECT count(temperature) as totalRows, sum(temperature) as sumTemperature, min(temperature) as minTemperature, max(temperature) as maxTemperature FROM readings where sensor_id = ? and date = ?`
  • 21. Application (Example) Total amount of data to scan: 1.44 billion points/day Coordinator Worker (loader machine) ScyllaDB cluster Worker (loader machine) Worker (loader machine) Set time frame, compute average, min, max of all sensors
  • 22. Packet is a global bare metal cloud that is built for Enterprises, and loved by Developers. TLDR: We sell servers to millennials.
  • 23. Core Edge Singapore Sydney Seattle Sunnyvale Dallas Amsterdam Los Angeles New York Tokyo Ashburn Atlanta Toronto Frankfurt Chicago Hong Kong Marseille Dublin 22+ Locations + Custom Deployments
  • 24. The Database (83 Machines) The Workers (24 Machines) Deployed via Packet’s API using Ansible playbooks.
  • 25. 83 Database Servers 2800 Physical Cores Intel Cascale Lake 314 TB of NVMe 34 TB of RAM
  • 26. Results Scanning 3 months of data Finished Scanning. Succeeded 132,480,000,000 rows. Failed 0 rows. RPC Failures: 0. Took 110,892.91 ms Processed 1,194,666,083 rows/s Absolute min: 19.71, date 2019-08-27, sensorID 473869 Absolute max: 135.21, date 2019-08-27, sensorID 473869
  • 27. Results Scanning 3 months of data Finished Scanning. Succeeded 132,480,000,000 rows. Failed 0 rows. RPC Failures: 0. Took 110,892.91 ms Processed 1,194,666,083 rows/s Absolute min: 19.71, date 2019-08-27, sensorID 473869 Absolute max: 135.21, date 2019-08-27, sensorID 473869
  • 28. Results Scanning 3 months of data Finished Scanning. Succeeded 132,480,000,000 rows. Failed 0 rows. RPC Failures: 0. Took 110,892.91 ms Processed 1,194,666,083 rows/s Absolute min: 19.71, date 2019-08-27, sensorID 473869 Absolute max: 135.21, date 2019-08-27, sensorID 473869
  • 29. Results Scanning the entire dataset Finished Scanning. Succeeded 525,599,474,400 rows. Failed 0 rows. RPC Failures: 0. Took 542,191.31 ms Processed 969,398,554 rows/s Absolute min: 68.00, date 2019-05-28, sensorID 82114 Absolute max: 79.99, date 2019-03-19, sensorID 152594
  • 30. Results Scanning the entire dataset Finished Scanning. Succeeded 525,599,474,400 rows. Failed 0 rows. RPC Failures: 0. Took 542,191.31 ms Processed 969,398,554 rows/s Absolute min: 68.00, date 2019-05-28, sensorID 82114 Absolute max: 79.99, date 2019-03-19, sensorID 152594
  • 31. Results Scanning the entire dataset Finished Scanning. Succeeded 525,599,474,400 rows. Failed 0 rows. RPC Failures: 0. Took 542,191.31 ms Processed 969,398,554 rows/s Absolute min: 68.00, date 2019-05-28, sensorID 82114 Absolute max: 79.99, date 2019-03-19, sensorID 152594
  • 33. We can efficiently process over 1.2 billion points per second (we’ll process whatever you need, too!)
  • 34. Bare metal + Scylla are like Peanut Butter & Jelly. + Packet & ScyllaDB together, provide a scalable & low latency database, deployed on a global bare metal cloud!
  • 35. “All Companies are Software Companies” Satya Nadella
  • 36. “All Companies are Software”
  • 37. “All Companies are Software Companies” Jay Kreps
  • 38. “All Companies are Software”
  • 39. “All Companies are Software”
  • 40. “All Companies are Software”
  • 41. Make it EASY for software companies
  • 42. Make it EASY for software companies ■ Scylla Cloud
  • 43. Make it EASY for software companies ■ Scylla Cloud ■ Project Alternator Scylla gains DynamoDB compatible API
  • 44. ■ Scylla Cloud ■ Project Alternator Scylla gains DynamoDB compatible API Make it EASY for software companies 120 Kops/sec is pretty intensive - in DynamoDB provisioned pricing, it would cost $85 per hour. VMs (EC2, on-demand pricing) for running Scylla - $7.5 per hour. Year reservation cheaper for both.
  • 45. Make it EASY for software companies Scylla Cloud Workload prioritization Scylla Manager BackupAlternator
  • 46. Make it EASY for software companies LWT Scylla Cloud Workload prioritization Scylla Manager BackupAlternator
  • 47. Make it EASY for software companies LWT Scylla Cloud Workload prioritization Scylla Manager BackupAlternator CDC
  • 48. Make it EASY for software companies UDF/ TransformationsLWT Scylla Cloud Workload prioritization Scylla Manager BackupAlternator CDC
  • 49. Make it EASY for software companies
  • 50. Scylla & ScyllaDB ■ Major Round of Funding ■ 2x Engineers ■ 2x Customers ■ $ you spend goes back to you
  • 51. Func6 Customer story ■ Required by a new customer ■ Although code ready, it wasn’t yet tested ■ We had to double down to release it
  • 52. Func6 Customer story Greg, account executive Eyal coordinated the requirement Shlomi, assigned people Luis, tested the OSS version Pekka built + back ported to enterprise Calle fixed bugs Gleb reviewed code Kamil reviewed & tested Michal designed, Fixed, built Scylla Manager Amnon adjusted the monitor Amos tested Roy tested Bentsi tested Julia tested
  • 53. “All Companies are Software Companies” “Software is Eating the World” “All Companies are Software”
  • 54. Scylla is a damn good Choice for *your* Software
  • 56. Where are we going Backup - point in time, auto-pre-upgrade Tablets k8s/manager
  • 57. Release Schedule 2nd Quarter 2020 3rd Quarter 2020 Apr May Jun Jul Aug Sep 4th Quarter 2019 Oct Nov Dec 1st Quarter 2020 Jan Feb Mar Scylla Enterprise 2019.1.3 Scylla Manager 2.0 Scylla 3.2 Scylla Enterprise 2019.1.4 Scylla Enterprise 2020.1 Scylla 4.0 Scylla Manager 2.1 Scylla Manager 3.0

Editor's Notes

  • #3: Who knows what’s the C10K problem? You’re old! The C10k problem is the problem of optimising network sockets to handle a large number of clients at the same time The term was coined in 1999 by Dan Kegel,[3][4] https://en.wikipedia.org/wiki/C10k_problem
  • #4: If you knew what C10K is, you’re old! Who uses FTP these days (it’s not a question) Originally NGINX was developed to solve this! Nice to see idevelopers redeisgn software for better performance and not another layer of indirection. Now C10K isn’t a real problem today, as you may know with Scylla, we’re bound by the number of ports.
  • #5: In distributed computing, performance EQUAL cost, it’s a cost optimization. Segment defined the $10M engineering problem (I call it D10M) It affected the company’s margin and business results - when they approached VCs, the later complained about their margins. Note that 2 Years ago their problem was a $1M problem.. Segment is a good company and applied lots of good procedures. Put that in the context of your own scale and speed. https://segment.com/blog/the-10m-engineering-problem https://segment.com/blog/the-million-dollar-eng-problem/
  • #6: How you get to D10M, easy, all it takes is engineers and a cloud account, engineers + cloud == fire As easy as fire spread in California. Without efficient infrastructure, you’ll be translating your technical debt to finacial debt.
  • #7: Similarly, in the database world, once upon a time, there was a single mysql server.
  • #8: Scale on from there
  • #9: Keep on scaling - move to Nosql
  • #10: Hello scalable architecture
  • #11: And it grows
  • #12: And scale with ease
  • #13: And here you are at the D10M problem
  • #15: We wanted to test our limits and see what can be achieved with Scylla, 1 million ops is not a news no more.
  • #23: We empower companies to make infrastructure their competitive advantage with automated bare metal that can be deployed anywhere.
  • #40: Marc Andreessen penned his famous “Why Software Is Eating the World” followed by MSFT Satya Nadella.
  • #41: You can see this becomes a reality with Kylie Jenner’s 7 pp company, fulfilled by Shopify https://www.shopify.com/kylie
  • #42: Recently Confluent’s Jay Kreps made another observation that all companies are software. Indeed businesses today are very different.
  • #43: In the old days, in order to take a loan, you’d physically arrive to the bank, fill a form and then The manual process would kick in. The credit officer reviews the request, passes it on to the risk officier, ..Couple of days later, you’d get your approval.
  • #44: Today, businesses like next insurance, Sofi, Lemonade automate everything and Can provide you an answer within minutes, some need to provide an answer in milliseconds There are no officers but services, micro services
  • #45: These services run 247 and move lots of data around in a merid of microservices and databases. It’s all complex and data passes through the pipes far too many times.
  • #46: We at Scylla wish to make it easy for software companies to create better software
  • #47: Scylla Cloud DBaaS is a very good start. We launched it a year ago and it’s been in GA since March.
  • #48: Compatibility is important as well. 2 months ago we presented project alternator. The initial release was open source. A Scylla alternative for the widely used DynamodB at 10x the speed. TODAY we announce the availability of alternator on our cloud as beta.
  • #49: Uber-app like experience - one click to fantastic performance A simple 3 medium node cluster can do 120k ops. A similar DynamoDB Pricing costs $85/hour while the on demand VMs cost $7.5
  • #50: We’ve made these improvements, from the cloud, through workload prioritization and Scylla Manager automatic backups But what next?
  • #51: Today we announce the availability of Lightweight Transactions a long anticipated feature in Scylla. Tomorrow Kostja will present LWT. We look at LWT not as a regular feature but As a key building block ahead, such that will allow us to improve Scylla itself (tablets, schema mgmt) And also to open the door to variety of more use cases to you all
  • #52: Next is Change Data Capture. We’ve given good amount of thought and design And I’m super pleased to announce the availability of CDC in OSS. Once again, we like to help build software companies and simplify your stack. CDC will allow to think about tables and streams in an interchangeable way
  • #53: Last but not least, user defined functions is here. UDF opens the gate to much more. Think about ETL within the Scylla servers Themselves with the Scylla parallelism. We’re not there yet but this is where we’re going
  • #54: Think avout filtering your stream of changes on the server and apply actions on it.
  • #55: Not only Scylla as a database has made progress, also Scylla as a company. We closed a major round of funding. We do developers all hands in between the scylla summits.
  • #56: We’re bigger and busier than ever but still put all of the emphasis in support. For example, one of our customers needed a functionality which was coded but not yet hardened. The contract was signed so we had to double down on it.
  • #58: Bob my good VP of marketing indicated that I’m quoting other people and whether I like to coin something myself. No sure what’s left for me, can’t shrink these sentences even More - software is software? Doesn’t make sense. Instead, I can say much less hyperbolic claim
  • #59: Scylla is damn good.. Hope you’ll enjoy the show!