Latency: the #1 metric of your cloud Boyan Krosnov Chief of - - PowerPoint PPT Presentation

▶
latency the 1 metric of your cloud boyan krosnov
SMART_READER_LITE
LIVE PREVIEW

Latency: the #1 metric of your cloud Boyan Krosnov Chief of - - PowerPoint PPT Presentation

May 10, 2023 447 likes •747 views

Latency: the #1 metric of your cloud Boyan Krosnov Chief of Product Cloud architect "External technology" Previously service providers, networks, packet processing (SDN) Algorithms & data structures 2 StorPool

slide-1
SLIDE 1

Latency: the #1 metric of your cloud

slide-2
SLIDE 2

Boyan Krosnov

  • Chief of Product
  • Cloud architect
  • "External technology"
  • Previously service providers, networks,

packet processing (SDN)

  • Algorithms & data structures

2

slide-3
SLIDE 3

StorPool

  • Fast and efficient software-defined storage

system

  • Used by Private clouds, MSPs, Cloud service

providers, Mobile/Web Apps

  • Started in 2011 to solve storage for cloud

service providers

  • Clean slate design - scale-out, API-controlled,

end-to-end data integrity, CoW

3

slide-4
SLIDE 4

StorPool

  • Most deployments are with KVM

○ some Xen, some VMWare, some Hyper-V ○ some bare metal / dedicated server

  • Deep integrations into OpenStack,

Kubernetes, CloudStack, OpenNebula, OnApp

4

slide-5
SLIDE 5

5

Latency

  • ps per second
slide-6
SLIDE 6

6

core core core core task

slide-7
SLIDE 7

7

core core core core task task

slide-8
SLIDE 8

8

core core core core task task task task

slide-9
SLIDE 9

9

core core core core task task task task task task

slide-10
SLIDE 10

10

Latency

  • ps per second

s a t u r a t i

  • n

p

  • i

n t elastic mode c

  • n

g e s t e d m

  • d

e

slide-11
SLIDE 11

11

Latency

  • ps per second

best service

slide-12
SLIDE 12

12

Latency

  • ps per second

best service lowest cost per delivered resource

slide-13
SLIDE 13

13

Latency

  • ps per second

best service lowest cost per delivered resource

  • nly pain

system throughput

slide-14
SLIDE 14

14

Latency

  • ps per second

best service lowest cost per delivered resource

  • nly pain

benchmarks

slide-15
SLIDE 15

15

slide-16
SLIDE 16

8 vCPUs, 16 GB RAM vdisk with 20k IOPS "fast" volume with approx 0.15 ms latency (QD 1) "slow" volume with 2 ms latency (QD 1) added with dm-delay in host Both volumes are on the same pool Both volumes measure 20k IOPS flat

16

slide-17
SLIDE 17

17

Latency

  • ps per second

20k IOPS

slide-18
SLIDE 18

18

Latency

  • ps per second

20k IOPS 2ms

slide-19
SLIDE 19

19

Latency

  • ps per second

20k IOPS 0.15ms

slide-20
SLIDE 20

8 vCPUs, 16 GB RAM vdisk with 20k IOPS pgbench --client=16 --jobs=8 \

  • -progress=1 --time=10 pgbench09

database size: 64 GB (4x RAM) https://wiki.postgresql.org/wiki/Pgbenchtesting

20

slide-21
SLIDE 21

8 vCPUs, 16 GB RAM vdisk with 20k IOPS 2ms storage latency 1600 TPS @ 8 ms in pgbench

  • > if we ask for 2500 TPS -> pile up

21

slide-22
SLIDE 22

8 vCPUs, 16 GB RAM vdisk with 20k IOPS 0.15ms storage latency 3600 TPS @ 3 ms in pgbench

  • > if we ask for 2500 TPS - no problem

22

slide-23
SLIDE 23

random read/write 50/50 QD 1, avg. latency

23

DigitalOcean (Ceph) 1.75 ms OVH (Ceph) 1.53 ms Tencent (?) 0.76 ms AWS gp2 10k 0.29 ms Alibaba (?) 0.18 ms StorPool BCP 0.17 ms

slide-24
SLIDE 24

random read/write 50/50 QD 1, avg. latency

24

DigitalOcean (Ceph) 1.75 ms OVH (Ceph) 1.53 ms Tencent (?) 0.76 ms AWS gp2 10k 0.29 ms Alibaba (?) 0.18 ms StorPool BCP 0.17 ms

slide-25
SLIDE 25

25 8 vCPUs, 16 GB RAM, dedicated if possible database size = 4x RAM https://wiki.postgresql.org/wiki/Pgbenchtesting

slide-26
SLIDE 26

Q&A

26

slide-27
SLIDE 27

Thank you!

27

slide-28
SLIDE 28

Boyan Krosnov StorPool Storage bk@storpool.com

www.storpool.com @storpool

28