In this paper we tackle the problem of providing Quality of Service guarantees to virtualized applications, focusing on computing and networking guarantees. We propose a mechanism for providing temporal isolation based on a CPU real time scheduling strategy. This allows not only to have control over the individual virtual machine throughput, but also on the activation latency and response-time by which virtualized software components react to external events. We show experimental results gathered on a real system validating the approach.

1. VHPC 2010
August 31st, Ischia, Italy
Providing Performance Guarantees to
Virtual Machines using Real-Time
Scheduling
Tommaso Cucinotta, Dhaval Giani,
Dario Faggioli, Fabio Checconi
Real-Time Systems Laboratory
Scuola Superiore Sant'Anna
Pisa, Italy
Tommaso Cucinotta – ReTiS Lab – Scuola Superiore Sant'Anna – Pisa – Italy 1/24

2. Introduction and Motivations
Tommaso Cucinotta – ReTiS Lab – Scuola Superiore Sant'Anna – Pisa – Italy 2/24

3. Introduction
“Traditional” worlds of computing
 General-Purpose Computing (GPC) and servers
 Low-cost, low parallelism degree equipment
 Focused on mixing batch and interactive workloads
 High-Performance Computing (HPC)
 High-cost, massively parallel and vector-based equipment
 Focused on batch computing, tightly coupled parallel tasks,
scientific applications
New trends
 Affordable many-core systems also for GPC
 Cloud-Computing: world-level scalability and replicability
 Use of high-performance hardware in CC applications
 Interest in mixing interactive/real-time and HPC workloads
Tommaso Cucinotta – ReTiS Lab – Scuola Superiore Sant'Anna – Pisa – Italy 3/24

4. Introduction
Virtualization is a key technology
 For IaaS providers (Cloud Computing)
 For server consolidation
Physical Host
Physical Host
OS
OS
...
Physical Host
Physical Host
OS
OS
Tommaso Cucinotta – ReTiS Lab – Scuola Superiore Sant'Anna – Pisa – Italy 4/24

5. Introduction
Virtualization is a key technology
 For IaaS providers (Cloud Computing)
 For server consolidation
Physical Host
Physical Host Physical Host
Physical Host
OS
OS VM/OS
VM/OS
... ...
Physical Host VM/OS
VM/OS
Physical Host
OS
OS
Tommaso Cucinotta – ReTiS Lab – Scuola Superiore Sant'Anna – Pisa – Italy 5/24

6. Need for Performance Isolation
Resource sharing
→ Temporal interference
Physical Host
Physical Host
VM
VM
VM
VM
VM
VM
VM
VM
Tommaso Cucinotta – ReTiS Lab – Scuola Superiore Sant'Anna – Pisa – Italy 6/24

7. Need for Performance Isolation
~30ms
Resource sharing VM Alone
→ Temporal interference
Physical Host
Physical Host
VM
VM
VM
VM
VM
VM
VM
VM
Tommaso Cucinotta – ReTiS Lab – Scuola Superiore Sant'Anna – Pisa – Italy 7/24

8. Need for Performance Isolation
~30ms
Resource sharing VM Alone
→ Temporal interference
Physical Host
Physical Host
VM
VM
VM
VM
~120ms
2 VMs
VM
VM
VM
VM
Tommaso Cucinotta – ReTiS Lab – Scuola Superiore Sant'Anna – Pisa – Italy 8/24

9. Possible Solution
Hardware replication and Physical Host
Physical Host
static partitioning VM
VM
 Computing
 Multi-core (1 core per VM)
VM
VM
 Networking
 Multiple network adapters
(1 network adapter per VM) VM
VM
 Multi-queue adapters
Drawbacks VM
VM
 Limitation of flexibility
 Under-utilization of resources
 e.g., with real-time/interactive workloads
(cloud computing)
Tommaso Cucinotta – ReTiS Lab – Scuola Superiore Sant'Anna – Pisa – Italy 9/24

10. Possible Solution
Another approach
 Let multiple VMs use the same resources
 Use proper resource scheduling strategies
For example
 Computing
 Xen credit-based and SEDF schedulers
 Networking
 QoS-aware protocols (IntServ, MPLS)
Tommaso Cucinotta – ReTiS Lab – Scuola Superiore Sant'Anna – Pisa – Italy 10/24

11. What is still missing ?
Most approaches
 Investigate on high-level load balancing techniques
 Without the necessary care for fine-grain resource scheduling
 Focus on fairness among multiple VMs
Only a few works
 Focus on providing precise QoS guarantees to VMs
 For example
 Gupta et al., “[...] Performance Isolation […] in Xen”
 Cherkasova et al., “Comparison of the 3 CPU Schedulers in Xen”
Our focus
 Network performance isolation among concurrent VMs
 Virtualization with a host OS
 Specifically, on Linux as host OS with KVM
Tommaso Cucinotta – ReTiS Lab – Scuola Superiore Sant'Anna – Pisa – Italy 11/24

12. Proposed Approach
Tommaso Cucinotta – ReTiS Lab – Scuola Superiore Sant'Anna – Pisa – Italy 12/24

13. Proposed Approach
Use of real-time CPU scheduling
 As basic mechanism for isolating VMs concurrently
running on the same CPU and core
IRMOS Real-Time Scheduler
 For the Linux kernel
 Provides hierarchical EDF/FP scheduling
 EDF-based resource reservations
– (Q, P): a budget Q is granted every period P
– Both a guarantee and a limitation
 FP scheduling within each EDF reservation
 Provides temporal isolation
Tommaso Cucinotta – ReTiS Lab – Scuola Superiore Sant'Anna – Pisa – Italy 13/24

14. What can we achieve ?
(by CPU real-time scheduling)
CPU real-time scheduling achieves
 Performance isolation of
compute-intensive VMs
 Shown in other papers
 What about network-intensive VMs ?
Problems
 On the host OS (Linux)
 network-intensive VMs impose a big interrupt workload
– difficult to quantify and keep under control
– steals CPU from reservations of other VMs
 How can we fix this ?
Tommaso Cucinotta – ReTiS Lab – Scuola Superiore Sant'Anna – Pisa – Italy 14/24

15. Preliminary Solution
Preliminary solution (subject of this paper)
 Temporal isolation of VMs by real-time scheduling
 Scheduling parameters tuned according to benchmarked figures
– Reservation period set according to responsiveness requirements
– Budget needed for computations, plus
– Budget needed for sustaining network traffic
 Budget over-provisioning
 According to the foreseen interference due to
network-intensive VMs
Tommaso Cucinotta – ReTiS Lab – Scuola Superiore Sant'Anna – Pisa – Italy 15/24

16. Experimental Results
Tommaso Cucinotta – ReTiS Lab – Scuola Superiore Sant'Anna – Pisa – Italy 16/24

17. Experimental Results
(Q6600 @ 2.4 GHz, 1-Gbit Card)
Achievable network throughput as a function of
the CPU share reserved to the VM
 Measured via iperf: almost linear, as expected
Tommaso Cucinotta – ReTiS Lab – Scuola Superiore Sant'Anna – Pisa – Italy 17/24

18. Experimental Results
(2 VMs on the same core)
Throughput as a function of its own budget
 13% drop when
 own reservation is 35% (from ~300 Mbps to ~260 Mbps)
 and reservation of other VM increased from 10% to 40%
Tommaso Cucinotta – ReTiS Lab – Scuola Superiore Sant'Anna – Pisa – Italy 18/24

19. Experimental Results
(2 VMs on the same core)
Throughput as a function of its own budget
 ~14% budget over-provisioning needed
 for sustaining a ~300 Mbps throughput
 when reservation of other VM increased from 10% to 40%
Tommaso Cucinotta – ReTiS Lab – Scuola Superiore Sant'Anna – Pisa – Italy 19/24

20. Experimental Results
(2 VMs on the same core)
Throughput as a function of the budget reserved
to the other interfering VM
 Drop due to compute-intensive interfering VM lower than
the one due to network-intensive interfering VM
Tommaso Cucinotta – ReTiS Lab – Scuola Superiore Sant'Anna – Pisa – Italy 20/24

21. Experimental Results
(application-level benchmark)
Download time for a 100 KB file from Apache
 Periodic download requests every 20ms
 Response-times may be kept much more stable
by real-time scheduling
Tommaso Cucinotta – ReTiS Lab – Scuola Superiore Sant'Anna – Pisa – Italy 21/24

22. Conclusions and Future Work
Tommaso Cucinotta – ReTiS Lab – Scuola Superiore Sant'Anna – Pisa – Italy 22/24

23. Conclusions and Future Work
Conclusions
 We showed how to achieve network
performance isolation by CPU
real-time scheduling
 Technique to be used jointly with
traffic-shaping techniques
Planned Future Work
 Experiment with PREEMPT_RT
 IRQs handled in kernel threads
 Investigate on task-level isolation of network traffic of
different VMs (via real-time scheduling)
 Make KVM a QoS-aware hypervisor
 Investigate on the use of Adaptive Reservations
Tommaso Cucinotta – ReTiS Lab – Scuola Superiore Sant'Anna – Pisa – Italy 23/24

24. Thanks for your attention!
Questions ?
Tommaso Cucinotta – ReTiS Lab – Scuola Superiore Sant'Anna – Pisa – Italy 24/24