autonomic resource provisioning for cloud-based software
DESCRIPTION
The Third National Conference on Cloud Computing and Commerce (NC4), for more information please refer to: http://computing.dcu.ie/~pjamshidi/PDF/SEAMS2014.pdfTRANSCRIPT
Autonomic Resource Provisioning for
Cloud-Based Software
Pooyan Jamshidi
Supervisor: Dr. Claus Pahl
In collaboration with: Dr. Aakash Ahmad
IC4- Irish Centre for Cloud Computing and Commerce
School of Computing, Dublin City University
The Third National Conference on Cloud
Computing and Commerce (NC4)
Dublin, Ireland
April 15, [email protected]
Elasticity
Measured
Service
On
demand
Service
Ubiquitous
Network
Access
Resource
Pooling
Selling point!
~50% = wasted hardware
Actual traffic
Problem 1: ~75% wasted capacityActual
demand
Problem 2:
customer lost
Really like this??
Capacity we can provision with RobusT2Scale
scale-up =bigger VMs
scale-out = more VMs
0 50 1000
500
1000
1500Big Spike
0 50 100100
200
300
400
500Dual Phase
0 50 1000
1000
2000Large Variations
0 50 1000
200
400
600Quickly Varying
0 50 1000
500
1000Slowly Varying
0 50 1000
500
1000Steep Tri Phase
Quantitative value=> requires deep knowledge
Qualitative threshold like “high” can be an alternative
0 10 20 30 40 50 60 70 80 90 100-500
0
500
1000
1500
2000
Time (seconds)
Num
ber
of
hits
Original data
betta=0.10, gamma=0.94, rmse=308.1565, rrse=0.79703
betta=0.27, gamma=0.94, rmse=209.7852, rrse=0.54504
betta=0.80, gamma=0.94, rmse=272.6285, rrse=0.70858
0 10 20 30 40 50 60 70 80 90 100-500
0
500
1000
1500
2000
Time (seconds)
Num
ber
of
hits
Original data
betta=0.10, gamma=0.94, rmse=308.1565, rrse=0.79703
betta=0.27, gamma=0.94, rmse=209.7852, rrse=0.54504
betta=0.80, gamma=0.94, rmse=272.6285, rrse=0.70858
SUT Criteria Big spike Dual phaseLarge
variations
Quickly
varying
Slowly
varying
Steep tri
phase
with
RobusT2Scale
𝑟𝑡95% 973ms 537ms 509ms 451ms 423ms 498ms
𝑣𝑚 3.2 3.8 5.1 5.3 3.7 3.9
with
overprovisioning
𝑟𝑡95% 354ms 411ms 395ms 446ms 371ms 491ms
𝑣𝑚 6 6 6 6 6 6
with under
provisioning
𝑟𝑡95% 1465ms 1832ms 1789ms 1594ms 1898ms 2194ms
𝑣𝑚 2 2 2 2 2 2
SLA with 𝒓𝒕𝟗𝟓 = 𝟔𝟎𝟎𝒎𝒔
0
0.02
0.04
0.06
0.08
0.1
alpha=0.1 alpha=0.5 alpha=0.9 alpha=1.0
Benefits
Limitations, open issues
& future directionsSome people simply don’t like auto-scaling! They rather believe in capacity planningOn Why I Don't Like Auto-Scaling in the Cloud
0
0.02
0.04
0.06
0.08
0.1
alpha=0.1 alpha=0.5 alpha=0.9 alpha=1.0
Challenge 1: ~75% wasted capacityActual
demand
Challenge 2:
customer lost
http://computing.dcu.ie/~pjamshidi/PDF/SEAMS2014.pdf
More Details?=
>