SlideShare a Scribd company logo
1 of 63
Download to read offline
Go Big or Go Home!
Does it still make sense to do Big Data
with Small nodes?
WEBINAR
2
Glauber Costa
Glauber Costa is a Principal Architect at ScyllaDB.
He shares his time between the engineering department
working on upcoming Scylla features and helping
customers succeed.
Before ScyllaDB, Glauber worked with Virtualization in the
Linux Kernel for 10 years with contributions ranging from
the Xen Hypervisor to all sorts of guest functionality and
containers.
3
+ Next-generation NoSQL database
+ Drop-in replacement for Cassandra
+ 10X the performance & low tail latency
+ Open source and enterprise editions
+ Founded by the creators of KVM hypervisor
+ HQs: Palo Alto, CA; Herzelia, Israel
About ScyllaDB
A long, long time ago...
4
A long, long time ago...
5
+ NoSQL allows Big Data with
commodity HW.
+ 2008: Intel Core, 2 cores.
+ 2018: Samsung S8, octacore,
+ fits in your pocket.
A long, long time ago...
6
+ NoSQL allows Big Data with
commodity HW.
+ 2008: Intel Core, 2 cores.
+ 2018: Samsung S8, octacore,
+ fits in your pocket.
+ Need to store 200TB of data:
+ 200 nodes, 4 cores and 1TB each, or:
+ 20 nodes, 40 cores and 10TB each?
7
Big vs Small?
+ Bigger nodes have fewer noisy neighbors.
+ Bigger nodes see economies of scale.
+ Fewer nodes increase manageability.
+ But I do small nodes because each of them only have 500GB of disk
anyway!
8
Commodity Hardware in 2018
9
Commodity Hardware in 2018
2x
10
Commodity Hardware in 2018
2x
11
Commodity Hardware in 2018
2x - most of the time
size and quantity.
12
Commodity Hardware in 2018
2x
13
+ MTBF is a constant. So twice the nodes means twice the failures.
+ Even assuming each individual failure takes more time to recover:
+ How many failures per year if 3-node cluster with 20TB each?
+ How many failures per year if 60-node cluster with 1TB each?
+ A part of the cost is per-failure, not per size:
+ How do you like being paged every week, instead of twice a year?
+ Security fix, kernel update: rolling update 3 nodes vs 60 nodes.
More nodes mean more failures
Find the inflection
point
14
Let’s do some experiments
1 c4.4xlarge, 250 threads, QUORUM writes 1,000,000,000 partitions
Cluster is 3 x i3.xlarge, RF=3
Let’s do some experiments
1 c4.4xlarge, 250 threads, QUORUM writes 1,000,000,000 partitions
Cluster is 3 x i3.xlarge, RF=3
latency mean : 12.9
Let’s do some experiments
1 c4.4xlarge, 250 threads, QUORUM writes 1,000,000,000 partitions
Cluster is 3 x i3.xlarge, RF=3
latency mean : 12.9
latency 95th percentile : 20.2
Let’s do some experiments
1 c4.4xlarge, 250 threads, QUORUM writes 1,000,000,000 partitions
Cluster is 3 x i3.xlarge, RF=3
latency mean : 12.9
latency 95th percentile : 20.2
latency 99th percentile : 26.2
Let’s do some experiments
1 c4.4xlarge, 250 threads, QUORUM writes 1,000,000,000 partitions
Cluster is 3 x i3.xlarge, RF=3
latency mean : 12.9
latency 95th percentile : 20.2
latency 99th percentile : 26.2
latency 99.9th percentile : 40.0
Let’s do some experiments
1 c4.4xlarge, 250 threads, QUORUM writes 1,000,000,000 partitions
Cluster is 3 x i3.xlarge, RF=3
latency mean : 12.9
latency 95th percentile : 20.2
latency 99th percentile : 26.2
latency 99.9th percentile : 40.0
Total operation time : 14:19:02
Let’s do some experiments
2 c4.4xlarge, 250 threads, QUORUM writes 2,000,000,000 partitions
Cluster is 3 x i3.2xlarge, RF=3 (max between all clients)
Let’s do some experiments
2 c4.4xlarge, 250 threads, QUORUM writes 2,000,000,000 partitions
Cluster is 3 x i3.2xlarge, RF=3 (max between all clients)
latency mean : 13.6
Let’s do some experiments
2 c4.4xlarge, 250 threads, QUORUM writes 2,000,000,000 partitions
Cluster is 3 x i3.2xlarge, RF=3 (max between all clients)
latency mean : 13.6
latency 95th percentile : 21.2
Let’s do some experiments
2 c4.4xlarge, 250 threads, QUORUM writes 2,000,000,000 partitions
Cluster is 3 x i3.2xlarge, RF=3 (max between all clients)
latency mean : 13.6
latency 95th percentile : 21.2
latency 99th percentile : 27.3
Let’s do some experiments
2 c4.4xlarge, 250 threads, QUORUM writes 2,000,000,000 partitions
Cluster is 3 x i3.2xlarge, RF=3 (max between all clients)
latency mean : 13.6
latency 95th percentile : 21.2
latency 99th percentile : 27.3
latency 99.9th percentile : 38.8
Let’s do some experiments
2 c4.4xlarge, 250 threads, QUORUM writes 2,000,000,000 partitions
Cluster is 3 x i3.2xlarge, RF=3 (max between all clients)
latency mean : 13.6
latency 95th percentile : 21.2
latency 99th percentile : 27.3
latency 99.9th percentile : 38.8
Total operation time : 15:09:49
Let’s do some experiments
2 c4.4xlarge, 250 threads, QUORUM writes 2,000,000,000 partitions
Cluster is 3 x i3.2xlarge, RF=3 (max between all clients)
latency mean : 13.6
latency 95th percentile : 21.2
latency 99th percentile : 27.3
latency 99.9th percentile : 38.8
Total operation time : 15:09:49 (+ 6%)
Let’s do some experiments
4 c4.4xlarge, 250 threads, QUORUM writes 4,000,000,000 partitions
Cluster is 3 x i3.4xlarge, RF=3 (max between all clients)
Let’s do some experiments
4 c4.4xlarge, 250 threads, QUORUM writes 4,000,000,000 partitions
Cluster is 3 x i3.4xlarge, RF=3 (max between all clients)
latency mean : 10.6
latency 95th percentile : 16.8
latency 99th percentile : 21.5
latency 99.9th percentile : 26.5
Let’s do some experiments
4 c4.4xlarge, 250 threads, QUORUM writes 4,000,000,000 partitions
Cluster is 3 x i3.4xlarge, RF=3 (max between all clients)
latency mean : 10.6
latency 95th percentile : 16.8
latency 99th percentile : 21.5
latency 99.9th percentile : 26.5
Total operation time : 11:44:26
Let’s do some experiments
4 c4.4xlarge, 250 threads, QUORUM writes 4,000,000,000 partitions
Cluster is 3 x i3.4xlarge, RF=3 (max between all clients)
latency mean : 10.6
latency 95th percentile : 16.8
latency 99th percentile : 21.5
latency 99.9th percentile : 26.5
Total operation time : 11:44:26 (- 22%)
Let’s do some experiments
8 c4.4xlarge, 250 threads, QUORUM writes 8,000,000,000 partitions
Cluster is 3 x i3.8xlarge, RF=3 (max between all clients)
Let’s do some experiments
8 c4.4xlarge, 250 threads, QUORUM writes 8,000,000,000 partitions
Cluster is 3 x i3.8xlarge, RF=3 (max between all clients)
11:48:11 (+ 5%)
Let’s do some experiments
16 c4.4xlarge, 250 threads, QUORUM writes 16,000,000,000 partitions
Cluster is 3 x i3.16xlarge, RF=3 (max between all clients)
Let’s do some experiments
16 c4.4xlarge, 250 threads, QUORUM writes 16,000,000,000 partitions
Cluster is 3 x i3.16xlarge, RF=3 (max between all clients)
12:30:04 (+ 6%)
Let’s do some experiments
Linear scale-up capabilities, as much as scale out. It pays to scale up
Total data size per node in the i3.16xlarge case is 4.8TB.
1B rows 2B rows 4B rows 8B rows 16B rows
time to ingest
Good luck
compacting that!
37
What is compaction again?
Writes
commit log
compaction
# concurrent_compactors: 1
#
# compaction_throughput_mb_per_sec: 16
All of those settings are ignored by Scylla
Handling Compactions
Query
Commitlog
Compaction
Queue
Queue
Queue
Userspace
I/O
Scheduler
Disk
Max useful disk concurrency
I/O queued in FS/deviceNo queues
Scylla I/O Scheduler
Let’s do some experiments
In one of the nodes from previous experiment:
nodetool compact from quiescent state, 293GB, i3.xlarge: 1:45:27
Let’s do some experiments
nodetool compact from quiescent state, 587GB i3.2xlarge: 1:47:05
Let’s do some experiments
nodetool compact from quiescent state, 1.2TB i3.4xlarge: 2:00:41
Let’s do some experiments
nodetool compact from quiescent state, 2.4TB i3.8xlarge: 2:02:59
Let’s do some experiments
nodetool compact from quiescent state, 4.8TB i3.16xlarge: 2:11:34
4.8TB2.4TB1.2TB0.6TB0.3TB
Time to fully compact the node
So what that the
node is down?
46
Heat-weighted Load Balancing
+ Replica goes down and comes back up
+ Caches are cold.
+ Never sending requests to the node means caches never warm up.
+ Optimize mathematically the desired hit ratio so that caches warm up,
+ while keeping the latencies down
47
Restart
Heat-weighted Load Balancing
48
Heat-weighted Load Balancing
49
Restarted node. Cache misses are initially
high but deterministically go down
Heat-weighted Load Balancing
50
majority of reads are
sent to other replicas
Heat-weighted Load Balancing
51
majority of reads are
sent to other replicas
until cache is warm at the
restarted node
Heat-weighted Load Balancing
52
Heat-weighted Load Balancing
53
No-HWLB: high latencies
on restart until warm
Heat-weighted Load Balancing
54
No-HWLB: high latencies
on restart until warm
(100s of ms, 99 %)
HWLB: 99 % is in single-
digit ms, few 10s ms peaks
Heat-weighted Load Balancing
55
+ Before
+ After
Conclusion
56
+ Scylla scales linearly in the amount of resources
+ Linear scalability also reflects on compaction performance
+ During failures, features like Heat Weighted Load Balance help the cluster
keep SLAs
+ so the fact that it takes longer to recover is not an issue
But wait:
Is it true?
57
# stream_throughput_outbound_megabits_per_sec: 200
+ Transfer 4.8 TB over 25Gbps @linerate: ~1500s,
+ We are pretty much not hitting a network bottleneck.
Is the network a bottleneck?
The real cost of streaming
59
+ Same clusters as previous experiments.
+ Destroy compacted node, rebuild from remaining two.
The real cost of streaming
60
+ Same clusters as previous experiments.
+ Destroy compacted node, rebuild from remaining two.
1B rows 2B rows 4B rows 8B rows 16B rows
4.8TB2.4TB1.2TB0.6TB0.3TB
Conclusion (revised)
61
+ Scylla scales linearly in the amount of resources
+ Linear scalability also reflects on compaction performance
+ During recovery, other features like Heat Weighted Load Balancing help keep
SLAs
+ so the fact that it takes longer to recover is not an issue
+ Larger nodes are not more expensive to recover on failures. That’s a myth.
62
glauber@scylladb.com
@glcst
@ScyllaDB
Stay in touch
Q&A
Learn more
Blog: scylladb.com/blog
United States
1900 Embarcadero Road
Palo Alto, CA 94303
Israel
11 Galgalei Haplada
Herzelia, Israel
www.scylladb.com
@scylladb
Thank You!

More Related Content

What's hot

Finding New Sub-Atomic Particles on the AWS Cloud (BDT402) | AWS re:Invent 2013
Finding New Sub-Atomic Particles on the AWS Cloud (BDT402) | AWS re:Invent 2013Finding New Sub-Atomic Particles on the AWS Cloud (BDT402) | AWS re:Invent 2013
Finding New Sub-Atomic Particles on the AWS Cloud (BDT402) | AWS re:Invent 2013Amazon Web Services
 
Speeding up R with Parallel Programming in the Cloud
Speeding up R with Parallel Programming in the CloudSpeeding up R with Parallel Programming in the Cloud
Speeding up R with Parallel Programming in the CloudRevolution Analytics
 
Ensuring Kubernetes Cost Efficiency across (many) Clusters - DevOps Gathering...
Ensuring Kubernetes Cost Efficiency across (many) Clusters - DevOps Gathering...Ensuring Kubernetes Cost Efficiency across (many) Clusters - DevOps Gathering...
Ensuring Kubernetes Cost Efficiency across (many) Clusters - DevOps Gathering...Henning Jacobs
 
AWS RDS Benchmark - CMG Brasil 2012
AWS RDS Benchmark - CMG Brasil 2012AWS RDS Benchmark - CMG Brasil 2012
AWS RDS Benchmark - CMG Brasil 2012Rodrigo Campos
 
PHP Backends for Real-Time User Interaction using Apache Storm.
PHP Backends for Real-Time User Interaction using Apache Storm.PHP Backends for Real-Time User Interaction using Apache Storm.
PHP Backends for Real-Time User Interaction using Apache Storm.DECK36
 
GPU cloud with Job scheduler and Container
GPU cloud with Job scheduler and ContainerGPU cloud with Job scheduler and Container
GPU cloud with Job scheduler and ContainerAndrew Yongjoon Kong
 
GRAPH 101- GETTING STARTED WITH TITAN AND CASSANDRA
GRAPH 101- GETTING STARTED WITH TITAN AND CASSANDRAGRAPH 101- GETTING STARTED WITH TITAN AND CASSANDRA
GRAPH 101- GETTING STARTED WITH TITAN AND CASSANDRAShaunak Das
 
Aws S3 uploading tricks 2016
Aws S3 uploading tricks 2016Aws S3 uploading tricks 2016
Aws S3 uploading tricks 2016Bogdan Naydenov
 
Real-Time Analytics with Kafka, Cassandra and Storm
Real-Time Analytics with Kafka, Cassandra and StormReal-Time Analytics with Kafka, Cassandra and Storm
Real-Time Analytics with Kafka, Cassandra and StormJohn Georgiadis
 
유연하고 확장성 있는 빅데이터 처리
유연하고 확장성 있는 빅데이터 처리유연하고 확장성 있는 빅데이터 처리
유연하고 확장성 있는 빅데이터 처리NAVER D2
 
What makes AWS invincible? from JAWS Days 2014
What makes AWS invincible? from JAWS Days 2014What makes AWS invincible? from JAWS Days 2014
What makes AWS invincible? from JAWS Days 2014Emma Haruka Iwao
 
What's new in Ansible 2.0
What's new in Ansible 2.0What's new in Ansible 2.0
What's new in Ansible 2.0Allan Denot
 
Running Kubernetes in Production: A Million Ways to Crash Your Cluster - Cont...
Running Kubernetes in Production: A Million Ways to Crash Your Cluster - Cont...Running Kubernetes in Production: A Million Ways to Crash Your Cluster - Cont...
Running Kubernetes in Production: A Million Ways to Crash Your Cluster - Cont...Henning Jacobs
 
Scaling Apache Storm (Hadoop Summit 2015)
Scaling Apache Storm (Hadoop Summit 2015)Scaling Apache Storm (Hadoop Summit 2015)
Scaling Apache Storm (Hadoop Summit 2015)Robert Evans
 
Optimizing Kubernetes Resource Requests/Limits for Cost-Efficiency and Latency
Optimizing Kubernetes Resource Requests/Limits for Cost-Efficiency and LatencyOptimizing Kubernetes Resource Requests/Limits for Cost-Efficiency and Latency
Optimizing Kubernetes Resource Requests/Limits for Cost-Efficiency and LatencyHenning Jacobs
 
Docker with BGP - OpenDNS
Docker with BGP - OpenDNSDocker with BGP - OpenDNS
Docker with BGP - OpenDNSbacongobbler
 
Slide #1:Introduction to Apache Storm
Slide #1:Introduction to Apache StormSlide #1:Introduction to Apache Storm
Slide #1:Introduction to Apache StormMd. Shamsur Rahim
 
Docker at OpenDNS
Docker at OpenDNSDocker at OpenDNS
Docker at OpenDNSOpenDNS
 

What's hot (20)

Finding New Sub-Atomic Particles on the AWS Cloud (BDT402) | AWS re:Invent 2013
Finding New Sub-Atomic Particles on the AWS Cloud (BDT402) | AWS re:Invent 2013Finding New Sub-Atomic Particles on the AWS Cloud (BDT402) | AWS re:Invent 2013
Finding New Sub-Atomic Particles on the AWS Cloud (BDT402) | AWS re:Invent 2013
 
Speeding up R with Parallel Programming in the Cloud
Speeding up R with Parallel Programming in the CloudSpeeding up R with Parallel Programming in the Cloud
Speeding up R with Parallel Programming in the Cloud
 
Ensuring Kubernetes Cost Efficiency across (many) Clusters - DevOps Gathering...
Ensuring Kubernetes Cost Efficiency across (many) Clusters - DevOps Gathering...Ensuring Kubernetes Cost Efficiency across (many) Clusters - DevOps Gathering...
Ensuring Kubernetes Cost Efficiency across (many) Clusters - DevOps Gathering...
 
AWS RDS Benchmark - CMG Brasil 2012
AWS RDS Benchmark - CMG Brasil 2012AWS RDS Benchmark - CMG Brasil 2012
AWS RDS Benchmark - CMG Brasil 2012
 
PHP Backends for Real-Time User Interaction using Apache Storm.
PHP Backends for Real-Time User Interaction using Apache Storm.PHP Backends for Real-Time User Interaction using Apache Storm.
PHP Backends for Real-Time User Interaction using Apache Storm.
 
GPU cloud with Job scheduler and Container
GPU cloud with Job scheduler and ContainerGPU cloud with Job scheduler and Container
GPU cloud with Job scheduler and Container
 
GRAPH 101- GETTING STARTED WITH TITAN AND CASSANDRA
GRAPH 101- GETTING STARTED WITH TITAN AND CASSANDRAGRAPH 101- GETTING STARTED WITH TITAN AND CASSANDRA
GRAPH 101- GETTING STARTED WITH TITAN AND CASSANDRA
 
Aws S3 uploading tricks 2016
Aws S3 uploading tricks 2016Aws S3 uploading tricks 2016
Aws S3 uploading tricks 2016
 
Real-Time Analytics with Kafka, Cassandra and Storm
Real-Time Analytics with Kafka, Cassandra and StormReal-Time Analytics with Kafka, Cassandra and Storm
Real-Time Analytics with Kafka, Cassandra and Storm
 
유연하고 확장성 있는 빅데이터 처리
유연하고 확장성 있는 빅데이터 처리유연하고 확장성 있는 빅데이터 처리
유연하고 확장성 있는 빅데이터 처리
 
What makes AWS invincible? from JAWS Days 2014
What makes AWS invincible? from JAWS Days 2014What makes AWS invincible? from JAWS Days 2014
What makes AWS invincible? from JAWS Days 2014
 
Chainer v4 and v5
Chainer v4 and v5Chainer v4 and v5
Chainer v4 and v5
 
What's new in Ansible 2.0
What's new in Ansible 2.0What's new in Ansible 2.0
What's new in Ansible 2.0
 
Running Kubernetes in Production: A Million Ways to Crash Your Cluster - Cont...
Running Kubernetes in Production: A Million Ways to Crash Your Cluster - Cont...Running Kubernetes in Production: A Million Ways to Crash Your Cluster - Cont...
Running Kubernetes in Production: A Million Ways to Crash Your Cluster - Cont...
 
Scaling Apache Storm (Hadoop Summit 2015)
Scaling Apache Storm (Hadoop Summit 2015)Scaling Apache Storm (Hadoop Summit 2015)
Scaling Apache Storm (Hadoop Summit 2015)
 
Storm
StormStorm
Storm
 
Optimizing Kubernetes Resource Requests/Limits for Cost-Efficiency and Latency
Optimizing Kubernetes Resource Requests/Limits for Cost-Efficiency and LatencyOptimizing Kubernetes Resource Requests/Limits for Cost-Efficiency and Latency
Optimizing Kubernetes Resource Requests/Limits for Cost-Efficiency and Latency
 
Docker with BGP - OpenDNS
Docker with BGP - OpenDNSDocker with BGP - OpenDNS
Docker with BGP - OpenDNS
 
Slide #1:Introduction to Apache Storm
Slide #1:Introduction to Apache StormSlide #1:Introduction to Apache Storm
Slide #1:Introduction to Apache Storm
 
Docker at OpenDNS
Docker at OpenDNSDocker at OpenDNS
Docker at OpenDNS
 

Similar to Webinar: Does it Still Make Sense to do Big Data with Small Nodes?

Deep Dive on Amazon EC2 Instances (March 2017)
Deep Dive on Amazon EC2 Instances (March 2017)Deep Dive on Amazon EC2 Instances (March 2017)
Deep Dive on Amazon EC2 Instances (March 2017)Julien SIMON
 
Bulletproof Kafka with Fault Tree Analysis (Andrey Falko, Lyft) Kafka Summit ...
Bulletproof Kafka with Fault Tree Analysis (Andrey Falko, Lyft) Kafka Summit ...Bulletproof Kafka with Fault Tree Analysis (Andrey Falko, Lyft) Kafka Summit ...
Bulletproof Kafka with Fault Tree Analysis (Andrey Falko, Lyft) Kafka Summit ...confluent
 
AWS re:Invent 2016: [JK REPEAT] Deep Dive on Amazon EC2 Instances, Featuring ...
AWS re:Invent 2016: [JK REPEAT] Deep Dive on Amazon EC2 Instances, Featuring ...AWS re:Invent 2016: [JK REPEAT] Deep Dive on Amazon EC2 Instances, Featuring ...
AWS re:Invent 2016: [JK REPEAT] Deep Dive on Amazon EC2 Instances, Featuring ...Amazon Web Services
 
AWS re:Invent 2016: Deep Dive on Amazon EC2 Instances, Featuring Performance ...
AWS re:Invent 2016: Deep Dive on Amazon EC2 Instances, Featuring Performance ...AWS re:Invent 2016: Deep Dive on Amazon EC2 Instances, Featuring Performance ...
AWS re:Invent 2016: Deep Dive on Amazon EC2 Instances, Featuring Performance ...Amazon Web Services
 
Deep Dive on Amazon EC2 instances
Deep Dive on Amazon EC2 instancesDeep Dive on Amazon EC2 instances
Deep Dive on Amazon EC2 instancesAmazon Web Services
 
SRV402 Deep Dive on Amazon EC2 Instances, Featuring Performance Optimization ...
SRV402 Deep Dive on Amazon EC2 Instances, Featuring Performance Optimization ...SRV402 Deep Dive on Amazon EC2 Instances, Featuring Performance Optimization ...
SRV402 Deep Dive on Amazon EC2 Instances, Featuring Performance Optimization ...Amazon Web Services
 
The Best and Worst of Cassandra-stress Tool (Christopher Batey, The Last Pick...
The Best and Worst of Cassandra-stress Tool (Christopher Batey, The Last Pick...The Best and Worst of Cassandra-stress Tool (Christopher Batey, The Last Pick...
The Best and Worst of Cassandra-stress Tool (Christopher Batey, The Last Pick...DataStax
 
Comparing Apache Cassandra 4.0, 3.0, and ScyllaDB
Comparing Apache Cassandra 4.0, 3.0, and ScyllaDBComparing Apache Cassandra 4.0, 3.0, and ScyllaDB
Comparing Apache Cassandra 4.0, 3.0, and ScyllaDBScyllaDB
 
Accelerate Reed-Solomon coding for Fault-Tolerance in RAID-like system
Accelerate Reed-Solomon coding for Fault-Tolerance in RAID-like systemAccelerate Reed-Solomon coding for Fault-Tolerance in RAID-like system
Accelerate Reed-Solomon coding for Fault-Tolerance in RAID-like systemShuai Yuan
 
SRV402 Deep Dive on Amazon EC2 Instances, Featuring Performance Optimization ...
SRV402 Deep Dive on Amazon EC2 Instances, Featuring Performance Optimization ...SRV402 Deep Dive on Amazon EC2 Instances, Featuring Performance Optimization ...
SRV402 Deep Dive on Amazon EC2 Instances, Featuring Performance Optimization ...Amazon Web Services
 
SRV402 Deep Dive on Amazon EC2 Instances, Featuring Performance Optimization ...
SRV402 Deep Dive on Amazon EC2 Instances, Featuring Performance Optimization ...SRV402 Deep Dive on Amazon EC2 Instances, Featuring Performance Optimization ...
SRV402 Deep Dive on Amazon EC2 Instances, Featuring Performance Optimization ...Amazon Web Services
 
Accelerate Machine Learning on Google Cloud
Accelerate Machine Learning on Google CloudAccelerate Machine Learning on Google Cloud
Accelerate Machine Learning on Google CloudSamantha Guerriero
 
Alex Smola, Professor in the Machine Learning Department, Carnegie Mellon Uni...
Alex Smola, Professor in the Machine Learning Department, Carnegie Mellon Uni...Alex Smola, Professor in the Machine Learning Department, Carnegie Mellon Uni...
Alex Smola, Professor in the Machine Learning Department, Carnegie Mellon Uni...MLconf
 
Hs java open_party
Hs java open_partyHs java open_party
Hs java open_partyOpen Party
 
Production Ready Kafka on Kubernetes (Devandra Tagare, Lyft) Kafka Summit SF ...
Production Ready Kafka on Kubernetes (Devandra Tagare, Lyft) Kafka Summit SF ...Production Ready Kafka on Kubernetes (Devandra Tagare, Lyft) Kafka Summit SF ...
Production Ready Kafka on Kubernetes (Devandra Tagare, Lyft) Kafka Summit SF ...confluent
 
IMCSummit 2015 - Day 2 IT Business Track - 4 Myths about In-Memory Databases ...
IMCSummit 2015 - Day 2 IT Business Track - 4 Myths about In-Memory Databases ...IMCSummit 2015 - Day 2 IT Business Track - 4 Myths about In-Memory Databases ...
IMCSummit 2015 - Day 2 IT Business Track - 4 Myths about In-Memory Databases ...In-Memory Computing Summit
 
The Do’s and Don’ts of Benchmarking Databases
The Do’s and Don’ts of Benchmarking DatabasesThe Do’s and Don’ts of Benchmarking Databases
The Do’s and Don’ts of Benchmarking DatabasesScyllaDB
 
Mininet: Moving Forward
Mininet: Moving ForwardMininet: Moving Forward
Mininet: Moving ForwardON.Lab
 

Similar to Webinar: Does it Still Make Sense to do Big Data with Small Nodes? (20)

Deep Dive on Amazon EC2 Instances (March 2017)
Deep Dive on Amazon EC2 Instances (March 2017)Deep Dive on Amazon EC2 Instances (March 2017)
Deep Dive on Amazon EC2 Instances (March 2017)
 
Bulletproof Kafka with Fault Tree Analysis (Andrey Falko, Lyft) Kafka Summit ...
Bulletproof Kafka with Fault Tree Analysis (Andrey Falko, Lyft) Kafka Summit ...Bulletproof Kafka with Fault Tree Analysis (Andrey Falko, Lyft) Kafka Summit ...
Bulletproof Kafka with Fault Tree Analysis (Andrey Falko, Lyft) Kafka Summit ...
 
AWS re:Invent 2016: [JK REPEAT] Deep Dive on Amazon EC2 Instances, Featuring ...
AWS re:Invent 2016: [JK REPEAT] Deep Dive on Amazon EC2 Instances, Featuring ...AWS re:Invent 2016: [JK REPEAT] Deep Dive on Amazon EC2 Instances, Featuring ...
AWS re:Invent 2016: [JK REPEAT] Deep Dive on Amazon EC2 Instances, Featuring ...
 
AWS re:Invent 2016: Deep Dive on Amazon EC2 Instances, Featuring Performance ...
AWS re:Invent 2016: Deep Dive on Amazon EC2 Instances, Featuring Performance ...AWS re:Invent 2016: Deep Dive on Amazon EC2 Instances, Featuring Performance ...
AWS re:Invent 2016: Deep Dive on Amazon EC2 Instances, Featuring Performance ...
 
Docker vs kvm
Docker vs kvmDocker vs kvm
Docker vs kvm
 
Deep Dive on Amazon EC2 instances
Deep Dive on Amazon EC2 instancesDeep Dive on Amazon EC2 instances
Deep Dive on Amazon EC2 instances
 
SRV402 Deep Dive on Amazon EC2 Instances, Featuring Performance Optimization ...
SRV402 Deep Dive on Amazon EC2 Instances, Featuring Performance Optimization ...SRV402 Deep Dive on Amazon EC2 Instances, Featuring Performance Optimization ...
SRV402 Deep Dive on Amazon EC2 Instances, Featuring Performance Optimization ...
 
The Best and Worst of Cassandra-stress Tool (Christopher Batey, The Last Pick...
The Best and Worst of Cassandra-stress Tool (Christopher Batey, The Last Pick...The Best and Worst of Cassandra-stress Tool (Christopher Batey, The Last Pick...
The Best and Worst of Cassandra-stress Tool (Christopher Batey, The Last Pick...
 
Comparing Apache Cassandra 4.0, 3.0, and ScyllaDB
Comparing Apache Cassandra 4.0, 3.0, and ScyllaDBComparing Apache Cassandra 4.0, 3.0, and ScyllaDB
Comparing Apache Cassandra 4.0, 3.0, and ScyllaDB
 
Accelerate Reed-Solomon coding for Fault-Tolerance in RAID-like system
Accelerate Reed-Solomon coding for Fault-Tolerance in RAID-like systemAccelerate Reed-Solomon coding for Fault-Tolerance in RAID-like system
Accelerate Reed-Solomon coding for Fault-Tolerance in RAID-like system
 
SRV402 Deep Dive on Amazon EC2 Instances, Featuring Performance Optimization ...
SRV402 Deep Dive on Amazon EC2 Instances, Featuring Performance Optimization ...SRV402 Deep Dive on Amazon EC2 Instances, Featuring Performance Optimization ...
SRV402 Deep Dive on Amazon EC2 Instances, Featuring Performance Optimization ...
 
SRV402 Deep Dive on Amazon EC2 Instances, Featuring Performance Optimization ...
SRV402 Deep Dive on Amazon EC2 Instances, Featuring Performance Optimization ...SRV402 Deep Dive on Amazon EC2 Instances, Featuring Performance Optimization ...
SRV402 Deep Dive on Amazon EC2 Instances, Featuring Performance Optimization ...
 
Accelerate Machine Learning on Google Cloud
Accelerate Machine Learning on Google CloudAccelerate Machine Learning on Google Cloud
Accelerate Machine Learning on Google Cloud
 
Deep Dive on Amazon EC2
Deep Dive on Amazon EC2Deep Dive on Amazon EC2
Deep Dive on Amazon EC2
 
Alex Smola, Professor in the Machine Learning Department, Carnegie Mellon Uni...
Alex Smola, Professor in the Machine Learning Department, Carnegie Mellon Uni...Alex Smola, Professor in the Machine Learning Department, Carnegie Mellon Uni...
Alex Smola, Professor in the Machine Learning Department, Carnegie Mellon Uni...
 
Hs java open_party
Hs java open_partyHs java open_party
Hs java open_party
 
Production Ready Kafka on Kubernetes (Devandra Tagare, Lyft) Kafka Summit SF ...
Production Ready Kafka on Kubernetes (Devandra Tagare, Lyft) Kafka Summit SF ...Production Ready Kafka on Kubernetes (Devandra Tagare, Lyft) Kafka Summit SF ...
Production Ready Kafka on Kubernetes (Devandra Tagare, Lyft) Kafka Summit SF ...
 
IMCSummit 2015 - Day 2 IT Business Track - 4 Myths about In-Memory Databases ...
IMCSummit 2015 - Day 2 IT Business Track - 4 Myths about In-Memory Databases ...IMCSummit 2015 - Day 2 IT Business Track - 4 Myths about In-Memory Databases ...
IMCSummit 2015 - Day 2 IT Business Track - 4 Myths about In-Memory Databases ...
 
The Do’s and Don’ts of Benchmarking Databases
The Do’s and Don’ts of Benchmarking DatabasesThe Do’s and Don’ts of Benchmarking Databases
The Do’s and Don’ts of Benchmarking Databases
 
Mininet: Moving Forward
Mininet: Moving ForwardMininet: Moving Forward
Mininet: Moving Forward
 

Recently uploaded

Azure Monitor & Application Insight to monitor Infrastructure & Application
Azure Monitor & Application Insight to monitor Infrastructure & ApplicationAzure Monitor & Application Insight to monitor Infrastructure & Application
Azure Monitor & Application Insight to monitor Infrastructure & ApplicationAndikSusilo4
 
Scaling API-first – The story of a global engineering organization
Scaling API-first – The story of a global engineering organizationScaling API-first – The story of a global engineering organization
Scaling API-first – The story of a global engineering organizationRadu Cotescu
 
The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024Rafal Los
 
SIEMENS: RAPUNZEL – A Tale About Knowledge Graph
SIEMENS: RAPUNZEL – A Tale About Knowledge GraphSIEMENS: RAPUNZEL – A Tale About Knowledge Graph
SIEMENS: RAPUNZEL – A Tale About Knowledge GraphNeo4j
 
Install Stable Diffusion in windows machine
Install Stable Diffusion in windows machineInstall Stable Diffusion in windows machine
Install Stable Diffusion in windows machinePadma Pradeep
 
Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)Allon Mureinik
 
Tech-Forward - Achieving Business Readiness For Copilot in Microsoft 365
Tech-Forward - Achieving Business Readiness For Copilot in Microsoft 365Tech-Forward - Achieving Business Readiness For Copilot in Microsoft 365
Tech-Forward - Achieving Business Readiness For Copilot in Microsoft 3652toLead Limited
 
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024BookNet Canada
 
08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking Men08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking MenDelhi Call girls
 
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
04-2024-HHUG-Sales-and-Marketing-Alignment.pptxHampshireHUG
 
The Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptxThe Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptxMalak Abu Hammad
 
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024BookNet Canada
 
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...Patryk Bandurski
 
Human Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsHuman Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsMark Billinghurst
 
Understanding the Laravel MVC Architecture
Understanding the Laravel MVC ArchitectureUnderstanding the Laravel MVC Architecture
Understanding the Laravel MVC ArchitecturePixlogix Infotech
 
Key Features Of Token Development (1).pptx
Key  Features Of Token  Development (1).pptxKey  Features Of Token  Development (1).pptx
Key Features Of Token Development (1).pptxLBM Solutions
 
Factors to Consider When Choosing Accounts Payable Services Providers.pptx
Factors to Consider When Choosing Accounts Payable Services Providers.pptxFactors to Consider When Choosing Accounts Payable Services Providers.pptx
Factors to Consider When Choosing Accounts Payable Services Providers.pptxKatpro Technologies
 
Breaking the Kubernetes Kill Chain: Host Path Mount
Breaking the Kubernetes Kill Chain: Host Path MountBreaking the Kubernetes Kill Chain: Host Path Mount
Breaking the Kubernetes Kill Chain: Host Path MountPuma Security, LLC
 
Beyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
Beyond Boundaries: Leveraging No-Code Solutions for Industry InnovationBeyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
Beyond Boundaries: Leveraging No-Code Solutions for Industry InnovationSafe Software
 
Handwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed textsHandwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed textsMaria Levchenko
 

Recently uploaded (20)

Azure Monitor & Application Insight to monitor Infrastructure & Application
Azure Monitor & Application Insight to monitor Infrastructure & ApplicationAzure Monitor & Application Insight to monitor Infrastructure & Application
Azure Monitor & Application Insight to monitor Infrastructure & Application
 
Scaling API-first – The story of a global engineering organization
Scaling API-first – The story of a global engineering organizationScaling API-first – The story of a global engineering organization
Scaling API-first – The story of a global engineering organization
 
The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024
 
SIEMENS: RAPUNZEL – A Tale About Knowledge Graph
SIEMENS: RAPUNZEL – A Tale About Knowledge GraphSIEMENS: RAPUNZEL – A Tale About Knowledge Graph
SIEMENS: RAPUNZEL – A Tale About Knowledge Graph
 
Install Stable Diffusion in windows machine
Install Stable Diffusion in windows machineInstall Stable Diffusion in windows machine
Install Stable Diffusion in windows machine
 
Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)
 
Tech-Forward - Achieving Business Readiness For Copilot in Microsoft 365
Tech-Forward - Achieving Business Readiness For Copilot in Microsoft 365Tech-Forward - Achieving Business Readiness For Copilot in Microsoft 365
Tech-Forward - Achieving Business Readiness For Copilot in Microsoft 365
 
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
Transcript: #StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
 
08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking Men08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking Men
 
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
04-2024-HHUG-Sales-and-Marketing-Alignment.pptx
 
The Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptxThe Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptx
 
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
 
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
 
Human Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR SystemsHuman Factors of XR: Using Human Factors to Design XR Systems
Human Factors of XR: Using Human Factors to Design XR Systems
 
Understanding the Laravel MVC Architecture
Understanding the Laravel MVC ArchitectureUnderstanding the Laravel MVC Architecture
Understanding the Laravel MVC Architecture
 
Key Features Of Token Development (1).pptx
Key  Features Of Token  Development (1).pptxKey  Features Of Token  Development (1).pptx
Key Features Of Token Development (1).pptx
 
Factors to Consider When Choosing Accounts Payable Services Providers.pptx
Factors to Consider When Choosing Accounts Payable Services Providers.pptxFactors to Consider When Choosing Accounts Payable Services Providers.pptx
Factors to Consider When Choosing Accounts Payable Services Providers.pptx
 
Breaking the Kubernetes Kill Chain: Host Path Mount
Breaking the Kubernetes Kill Chain: Host Path MountBreaking the Kubernetes Kill Chain: Host Path Mount
Breaking the Kubernetes Kill Chain: Host Path Mount
 
Beyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
Beyond Boundaries: Leveraging No-Code Solutions for Industry InnovationBeyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
Beyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
 
Handwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed textsHandwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed texts
 

Webinar: Does it Still Make Sense to do Big Data with Small Nodes?

  • 1. Go Big or Go Home! Does it still make sense to do Big Data with Small nodes? WEBINAR
  • 2. 2 Glauber Costa Glauber Costa is a Principal Architect at ScyllaDB. He shares his time between the engineering department working on upcoming Scylla features and helping customers succeed. Before ScyllaDB, Glauber worked with Virtualization in the Linux Kernel for 10 years with contributions ranging from the Xen Hypervisor to all sorts of guest functionality and containers.
  • 3. 3 + Next-generation NoSQL database + Drop-in replacement for Cassandra + 10X the performance & low tail latency + Open source and enterprise editions + Founded by the creators of KVM hypervisor + HQs: Palo Alto, CA; Herzelia, Israel About ScyllaDB
  • 4. A long, long time ago... 4
  • 5. A long, long time ago... 5 + NoSQL allows Big Data with commodity HW. + 2008: Intel Core, 2 cores. + 2018: Samsung S8, octacore, + fits in your pocket.
  • 6. A long, long time ago... 6 + NoSQL allows Big Data with commodity HW. + 2008: Intel Core, 2 cores. + 2018: Samsung S8, octacore, + fits in your pocket. + Need to store 200TB of data: + 200 nodes, 4 cores and 1TB each, or: + 20 nodes, 40 cores and 10TB each?
  • 7. 7 Big vs Small? + Bigger nodes have fewer noisy neighbors. + Bigger nodes see economies of scale. + Fewer nodes increase manageability. + But I do small nodes because each of them only have 500GB of disk anyway!
  • 11. 11 Commodity Hardware in 2018 2x - most of the time size and quantity.
  • 13. 13 + MTBF is a constant. So twice the nodes means twice the failures. + Even assuming each individual failure takes more time to recover: + How many failures per year if 3-node cluster with 20TB each? + How many failures per year if 60-node cluster with 1TB each? + A part of the cost is per-failure, not per size: + How do you like being paged every week, instead of twice a year? + Security fix, kernel update: rolling update 3 nodes vs 60 nodes. More nodes mean more failures
  • 15. Let’s do some experiments 1 c4.4xlarge, 250 threads, QUORUM writes 1,000,000,000 partitions Cluster is 3 x i3.xlarge, RF=3
  • 16. Let’s do some experiments 1 c4.4xlarge, 250 threads, QUORUM writes 1,000,000,000 partitions Cluster is 3 x i3.xlarge, RF=3 latency mean : 12.9
  • 17. Let’s do some experiments 1 c4.4xlarge, 250 threads, QUORUM writes 1,000,000,000 partitions Cluster is 3 x i3.xlarge, RF=3 latency mean : 12.9 latency 95th percentile : 20.2
  • 18. Let’s do some experiments 1 c4.4xlarge, 250 threads, QUORUM writes 1,000,000,000 partitions Cluster is 3 x i3.xlarge, RF=3 latency mean : 12.9 latency 95th percentile : 20.2 latency 99th percentile : 26.2
  • 19. Let’s do some experiments 1 c4.4xlarge, 250 threads, QUORUM writes 1,000,000,000 partitions Cluster is 3 x i3.xlarge, RF=3 latency mean : 12.9 latency 95th percentile : 20.2 latency 99th percentile : 26.2 latency 99.9th percentile : 40.0
  • 20. Let’s do some experiments 1 c4.4xlarge, 250 threads, QUORUM writes 1,000,000,000 partitions Cluster is 3 x i3.xlarge, RF=3 latency mean : 12.9 latency 95th percentile : 20.2 latency 99th percentile : 26.2 latency 99.9th percentile : 40.0 Total operation time : 14:19:02
  • 21. Let’s do some experiments 2 c4.4xlarge, 250 threads, QUORUM writes 2,000,000,000 partitions Cluster is 3 x i3.2xlarge, RF=3 (max between all clients)
  • 22. Let’s do some experiments 2 c4.4xlarge, 250 threads, QUORUM writes 2,000,000,000 partitions Cluster is 3 x i3.2xlarge, RF=3 (max between all clients) latency mean : 13.6
  • 23. Let’s do some experiments 2 c4.4xlarge, 250 threads, QUORUM writes 2,000,000,000 partitions Cluster is 3 x i3.2xlarge, RF=3 (max between all clients) latency mean : 13.6 latency 95th percentile : 21.2
  • 24. Let’s do some experiments 2 c4.4xlarge, 250 threads, QUORUM writes 2,000,000,000 partitions Cluster is 3 x i3.2xlarge, RF=3 (max between all clients) latency mean : 13.6 latency 95th percentile : 21.2 latency 99th percentile : 27.3
  • 25. Let’s do some experiments 2 c4.4xlarge, 250 threads, QUORUM writes 2,000,000,000 partitions Cluster is 3 x i3.2xlarge, RF=3 (max between all clients) latency mean : 13.6 latency 95th percentile : 21.2 latency 99th percentile : 27.3 latency 99.9th percentile : 38.8
  • 26. Let’s do some experiments 2 c4.4xlarge, 250 threads, QUORUM writes 2,000,000,000 partitions Cluster is 3 x i3.2xlarge, RF=3 (max between all clients) latency mean : 13.6 latency 95th percentile : 21.2 latency 99th percentile : 27.3 latency 99.9th percentile : 38.8 Total operation time : 15:09:49
  • 27. Let’s do some experiments 2 c4.4xlarge, 250 threads, QUORUM writes 2,000,000,000 partitions Cluster is 3 x i3.2xlarge, RF=3 (max between all clients) latency mean : 13.6 latency 95th percentile : 21.2 latency 99th percentile : 27.3 latency 99.9th percentile : 38.8 Total operation time : 15:09:49 (+ 6%)
  • 28. Let’s do some experiments 4 c4.4xlarge, 250 threads, QUORUM writes 4,000,000,000 partitions Cluster is 3 x i3.4xlarge, RF=3 (max between all clients)
  • 29. Let’s do some experiments 4 c4.4xlarge, 250 threads, QUORUM writes 4,000,000,000 partitions Cluster is 3 x i3.4xlarge, RF=3 (max between all clients) latency mean : 10.6 latency 95th percentile : 16.8 latency 99th percentile : 21.5 latency 99.9th percentile : 26.5
  • 30. Let’s do some experiments 4 c4.4xlarge, 250 threads, QUORUM writes 4,000,000,000 partitions Cluster is 3 x i3.4xlarge, RF=3 (max between all clients) latency mean : 10.6 latency 95th percentile : 16.8 latency 99th percentile : 21.5 latency 99.9th percentile : 26.5 Total operation time : 11:44:26
  • 31. Let’s do some experiments 4 c4.4xlarge, 250 threads, QUORUM writes 4,000,000,000 partitions Cluster is 3 x i3.4xlarge, RF=3 (max between all clients) latency mean : 10.6 latency 95th percentile : 16.8 latency 99th percentile : 21.5 latency 99.9th percentile : 26.5 Total operation time : 11:44:26 (- 22%)
  • 32. Let’s do some experiments 8 c4.4xlarge, 250 threads, QUORUM writes 8,000,000,000 partitions Cluster is 3 x i3.8xlarge, RF=3 (max between all clients)
  • 33. Let’s do some experiments 8 c4.4xlarge, 250 threads, QUORUM writes 8,000,000,000 partitions Cluster is 3 x i3.8xlarge, RF=3 (max between all clients) 11:48:11 (+ 5%)
  • 34. Let’s do some experiments 16 c4.4xlarge, 250 threads, QUORUM writes 16,000,000,000 partitions Cluster is 3 x i3.16xlarge, RF=3 (max between all clients)
  • 35. Let’s do some experiments 16 c4.4xlarge, 250 threads, QUORUM writes 16,000,000,000 partitions Cluster is 3 x i3.16xlarge, RF=3 (max between all clients) 12:30:04 (+ 6%)
  • 36. Let’s do some experiments Linear scale-up capabilities, as much as scale out. It pays to scale up Total data size per node in the i3.16xlarge case is 4.8TB. 1B rows 2B rows 4B rows 8B rows 16B rows time to ingest
  • 38. What is compaction again? Writes commit log compaction
  • 39. # concurrent_compactors: 1 # # compaction_throughput_mb_per_sec: 16 All of those settings are ignored by Scylla Handling Compactions
  • 40. Query Commitlog Compaction Queue Queue Queue Userspace I/O Scheduler Disk Max useful disk concurrency I/O queued in FS/deviceNo queues Scylla I/O Scheduler
  • 41. Let’s do some experiments In one of the nodes from previous experiment: nodetool compact from quiescent state, 293GB, i3.xlarge: 1:45:27
  • 42. Let’s do some experiments nodetool compact from quiescent state, 587GB i3.2xlarge: 1:47:05
  • 43. Let’s do some experiments nodetool compact from quiescent state, 1.2TB i3.4xlarge: 2:00:41
  • 44. Let’s do some experiments nodetool compact from quiescent state, 2.4TB i3.8xlarge: 2:02:59
  • 45. Let’s do some experiments nodetool compact from quiescent state, 4.8TB i3.16xlarge: 2:11:34 4.8TB2.4TB1.2TB0.6TB0.3TB Time to fully compact the node
  • 46. So what that the node is down? 46
  • 47. Heat-weighted Load Balancing + Replica goes down and comes back up + Caches are cold. + Never sending requests to the node means caches never warm up. + Optimize mathematically the desired hit ratio so that caches warm up, + while keeping the latencies down 47 Restart
  • 49. Heat-weighted Load Balancing 49 Restarted node. Cache misses are initially high but deterministically go down
  • 50. Heat-weighted Load Balancing 50 majority of reads are sent to other replicas
  • 51. Heat-weighted Load Balancing 51 majority of reads are sent to other replicas until cache is warm at the restarted node
  • 53. Heat-weighted Load Balancing 53 No-HWLB: high latencies on restart until warm
  • 54. Heat-weighted Load Balancing 54 No-HWLB: high latencies on restart until warm (100s of ms, 99 %) HWLB: 99 % is in single- digit ms, few 10s ms peaks
  • 56. Conclusion 56 + Scylla scales linearly in the amount of resources + Linear scalability also reflects on compaction performance + During failures, features like Heat Weighted Load Balance help the cluster keep SLAs + so the fact that it takes longer to recover is not an issue
  • 57. But wait: Is it true? 57
  • 58. # stream_throughput_outbound_megabits_per_sec: 200 + Transfer 4.8 TB over 25Gbps @linerate: ~1500s, + We are pretty much not hitting a network bottleneck. Is the network a bottleneck?
  • 59. The real cost of streaming 59 + Same clusters as previous experiments. + Destroy compacted node, rebuild from remaining two.
  • 60. The real cost of streaming 60 + Same clusters as previous experiments. + Destroy compacted node, rebuild from remaining two. 1B rows 2B rows 4B rows 8B rows 16B rows 4.8TB2.4TB1.2TB0.6TB0.3TB
  • 61. Conclusion (revised) 61 + Scylla scales linearly in the amount of resources + Linear scalability also reflects on compaction performance + During recovery, other features like Heat Weighted Load Balancing help keep SLAs + so the fact that it takes longer to recover is not an issue + Larger nodes are not more expensive to recover on failures. That’s a myth.
  • 63. United States 1900 Embarcadero Road Palo Alto, CA 94303 Israel 11 Galgalei Haplada Herzelia, Israel www.scylladb.com @scylladb Thank You!