1、 ETSI GS NFV-INF 010 V1.1.1 (2014-12) Network Functions Virtualisation (NFV); Service Quality Metrics Disclaimer This document has been produced and approved by the Network Functions Virtualisation (NFV) ETSI Industry Specification Group (ISG) and represents the views of those members who participat
2、ed in this ISG. It does not necessarily represent the views of the entire ETSI membership. GROUP SPECIFICATION ETSI ETSI GS NFV-INF 010 V1.1.1 (2014-12) 2Reference DGS/NFV-INF010 Keywords measurement, NFV, quality ETSI 650 Route des Lucioles F-06921 Sophia Antipolis Cedex - FRANCE Tel.: +33 4 92 94
3、42 00 Fax: +33 4 93 65 47 16 Siret N 348 623 562 00017 - NAF 742 C Association but non lucratif enregistre la Sous-Prfecture de Grasse (06) N 7803/88 Important notice The present document can be downloaded from: http:/www.etsi.org The present document may be made available in electronic versions and
4、/or in print. The content of any electronic and/or print versions of the present document shall not be modified without the prior written authorization of ETSI. In case of any existing or perceived difference in contents between such versions and/or in print, the only prevailing document is the prin
5、t of the Portable Document Format (PDF) version kept on a specific network drive within ETSI Secretariat. Users of the present document should be aware that the document may be subject to revision or change of status. Information on the current status of this and other ETSI documents is available at
6、 http:/portal.etsi.org/tb/status/status.asp If you find errors in the present document, please send your comment to one of the following services: http:/portal.etsi.org/chaircor/ETSI_support.asp Copyright Notification No part may be reproduced or utilized in any form or by any means, electronic or m
7、echanical, including photocopying and microfilm except as authorized by written permission of ETSI. The content of the PDF version shall not be modified without the written authorization of ETSI. The copyright and the foregoing restriction extend to reproduction in all media. European Telecommunicat
8、ions Standards Institute 2014. All rights reserved. DECTTM, PLUGTESTSTM, UMTSTMand the ETSI logo are Trade Marks of ETSI registered for the benefit of its Members. 3GPPTM and LTE are Trade Marks of ETSI registered for the benefit of its Members and of the 3GPP Organizational Partners. GSM and the GS
9、M logo are Trade Marks registered and owned by the GSM Association. ETSI ETSI GS NFV-INF 010 V1.1.1 (2014-12) 3Contents Intellectual Property Rights 4g3Foreword . 4g3Modal verbs terminology 4g3Introduction 4g31 Scope 6g32 References 6g32.1 Normative references . 6g32.2 Informative references 6g33 De
10、finitions and abbreviations . 7g33.1 Definitions 7g33.2 Abbreviations . 8g34 NFV Service Quality Metrics Taxonomy. 9g35 Virtual Machine Service Quality Metrics. 11g35.1 VM Stall . 13g35.2 Premature VM Release Ratio . 13g35.3 VM Scheduling Latency . 14g35.4 VM Clock Error . 14g36 Virtual Network Inte
11、rface Service Quality Metrics . 15g37 Technology Component Service Quality Metrics 16g38 Orchestration Service Quality Metrics . 17g39 Service Quality Metrics Use Case 18g39.1 Short Term Use of Service Quality Metrics . 18g39.2 Medium Term Use of Service Quality Metrics 19g39.3 Long Term Use of Serv
12、ice Quality Metrics . 19g310 Recommendations 19g310.1 Measurement of Service Quality Metrics . 19g310.2 Service Quality Metrics in SLAs 20g310.3 Detailed Metric Definitions 20g310.4 Methods of Measurement . 20g310.5 Reporting Statistics and Results Processing . 20g310.6 Characterization Plans 20g3An
13、nex A (informative): Example NFVI-VIM Interactions related to Service Quality Metrics 21g3A.1 Introduction 21g3A.2 Resource Establishment (VM) . 21g3A.2.1 VM Provisioning Latency 21g3A.2.2 VM Provisioning Reliability 22g3A.2.3 VM DOA Ratio 22g3A.2.4 VM Placement Policy Compliance 23g3A.3 Resource Es
14、tablishment (Infrastructure Network) . 23g3A.3.1 VN Provisioning Latency . 24g3A.3.2 VN Provisioning Reliability . 24g3Annex B (informative): Authors Essential, or potentially Essential, IPRs notified to ETSI in respect of ETSI standards“, which is available from the ETSI Secretariat. Latest updates
15、 are available on the ETSI Web server (http:/ipr.etsi.org). Pursuant to the ETSI IPR Policy, no investigation, including IPR searches, has been carried out by ETSI. No guarantee can be given as to the existence of other IPRs not referenced in ETSI SR 000 314 (or the updates on the ETSI Web server) w
16、hich are, or may be, or may become, essential to the present document. Foreword This Group Specification (GS) has been produced by ETSI Industry Specification Group (ISG) Network Functions Virtualisation (NFV). The present document deals with specific aspects of Service Quality Metrics in the contex
17、t of Network Function Virtualisation. Infrastructure Architecture Document Document # Overview GS NFV INF 001 Illustrative Use Cases for the NFV Infrastructure GS NFV INF 002 Architecture of the Infrastructure Domains Compute Domain GS NFV INF 003 Hypervisor Domain GS NFV INF 004 Infrastructure Netw
18、ork Domain GS NFV INF 005 Architectural Methodology Interfaces and Abstraction GS NFV INF 007 Service Quality Metrics GS NFV INF 010 Modal verbs terminology In the present document “shall“, “shall not“, “should“, “should not“, “may“, “may not“, “need“, “need not“, “will“, “will not“, “can“ and “cann
19、ot“ are to be interpreted as described in clause 3.2 of the ETSI Drafting Rules (Verbal forms for the expression of provisions). “must“ and “must not“ are NOT allowed in ETSI deliverables except when used in direct citation. Introduction As shown in figure 1, the service quality delivered by a VNF i
20、nstance to end users is dependent on the service quality of the compute, network and other resources delivered by NFV infrastructure, VIM, VNFM and NFVO to the VNF instance. Objective and quantitative metrics for the service delivered by NFV infrastructure and orchestration to NFV consumers enables:
21、 Better engineering of VNF user service quality. Efficient fault localization and mitigation. Faster identification of true root cause of service impairment so proper corrective actions can be taken promptly. ETSI ETSI GS NFV-INF 010 V1.1.1 (2014-12) 5relies on acceptable NFV infrastructure service
22、being delivered to VNF componentsAssuring that VNF service quality delivered to users is acceptableQuantitative, objective metrics for service quality delivered by NFV cloud service provider to NFV consumers enables:checkbldbetter engineering of application (VNF) user service quality, checkbldeffici
23、ent fault localization and mitigation, and checkbldFaster identification of true root cause of service impairments so proper corrective actions can be taken promptlyFigure 1: Purpose of NFV Service Quality Metrics Figure 2 illustrates the service boundary characterized by these metrics between canon
24、ical NFV consumers who operate VNFs and NFVI service providers who operate NFV infrastructure and supporting systems. Objective metrics of the service quality delivered by the NFV infrastructure service provider to the NFV consumers VNFs enable quantitative discussions and agreements for the objecti
25、ves of NFV service quality to assure that the NFV consumers VNF instances deliver acceptable service quality to end users. NFVI (IaaS) Service ProviderNFV ConsumerNFV service quality metrics enable VNF suppliers, NFV consumers and NFV IaaSservice providers to agree on cloud service quality objective
26、s that enable VNFs to deliver acceptable service to the NFV consumers end usersEnd UsersFigure 2: Service Quality Metrics in the Context of NFV Reference Architecture Both new and familiar metrics will be needed to complete the quantification of service quality. The present document examines the pro
27、perties of the virtual infrastructure and describes metrics relevant and useful to virtualised functions. ETSI ETSI GS NFV-INF 010 V1.1.1 (2014-12) 61 Scope The present document enumerates metrics for NFV infrastructure, management and orchestration service qualities that can impact the end user ser
28、vice qualities delivered by VNF instances hosted on NFV infrastructure. These service quality metrics cover both direct service impairments, such as IP packets lost by NFV virtual networking which impacts end user service latency or quality of experience, and indirect service quality risks, such as
29、NFV management and orchestration failing to continuously and rigorously enforce all anti-affinity rules which increases the risk of an infrastructure failure causing unacceptable VNF user service impact. Performance relationships exist between the metrics described in this document and in other spec
30、ifications such as i.5. The present document does not consider: 1) Units of measurement for reporting, such as whether VM premature release rates should be expressed as hourly rate (e.g. 0,0001 premature VM release events per hour), annualized rate (e.g. 0,88 premature VM release events per year), h
31、ours between events (e.g. 10 000 hour mean time between premature release events), or events per other unit of time (e.g. 100 000 FITs, meaning 100 000 premature release events in one billion hours of operation). 2) Methods of Measurement which stipulate exactly how metrics will be measured. 3) Rigo
32、rous counting and exclusion rules, like the precise details given in the TL 9000 Measurements Handbook i.13. 4) Metrics that do not directly or indirectly impact VNF user service quality, like power efficiency. 2 References 2.1 Normative references References are either specific (identified by date
33、of publication and/or edition number or version number) or non-specific. For specific references, only the cited version applies. For non-specific references, the latest version of the referenced document (including any amendments) applies. Referenced documents which are not found to be publicly ava
34、ilable in the expected location might be found at http:/docbox.etsi.org/Reference. NOTE: While any hyperlinks included in this clause were valid at the time of publication, ETSI cannot guarantee their long term validity. The following referenced documents are necessary for the application of the pre
35、sent document. 1 ETSI GS NFV-INF 001: “Network Functions Virtualisation (NFV); Infrastructure Overview“. 2.2 Informative references References are either specific (identified by date of publication and/or edition number or version number) or non-specific. For specific references, only the cited vers
36、ion applies. For non-specific references, the latest version of the referenced document (including any amendments) applies. NOTE: While any hyperlinks included in this clause were valid at the time of publication, ETSI cannot guarantee their long term validity. The following referenced documents are
37、 not necessary for the application of the present document but they assist the user with regard to a particular subject area. i.1 NIST Special Publication 800-145 (September 2011): “The NIST Definitions of Cloud Computing“, Peter Mell and Timothy Grance, US National Institute of Standards and Techno
38、logy. i.2 IETF RFC 2330: “Framework for IP Performance Metrics“. ETSI ETSI GS NFV-INF 010 V1.1.1 (2014-12) 7i.3 Wiley-IEEE Press, 2013: “Service Quality of Cloud-Based Applications,“ Eric Bauer and Randee Adams. i.4 ETSI GS NFV-MAN 01: “Network Functions Virtualisation (NFV); Management and Orchestr
39、ation“. i.5 draft-ietf-ippm-model-based-metrics-02 (work in progress) (February 2014): “Model Based Bulk Performance Metrics“, M. Mathis and A. Morton. i.6 ETSI GS NFV-PER 001: “Network Functions Virtualisation (NFV); NFV Performance alternately, the VM can be prematurely released due to failure or
40、for other reasons (e.g. executing a lawful takedown order or non-payment of bill). ETSI ETSI GS NFV-INF 010 V1.1.1 (2014-12) 12Useful LifeTimeFailureIntensityINTEGRATION - T=0 is when new VNFC is fully operational and enters service in VNF.ALLOCATION - DOA Window begins when NFVI returns nominally w
41、orking VM instance to VNFVM ProvisioningREQUEST -VM provisioning request to NFVI initiated explicit action by NFV orchestrator or other entity VMDOAWindowRELEASE - VM useful life ends with orderly or premature (disorderly) VM release event.VM Useful Life WindowFigure 4: Virtual Machine Lifecycle VM-
42、related NFV service quality metrics before “T=0“ when the VM enters service (i.e. VM Provisioning Latency, VM Provisioning Reliability, VM DOA Ratio) are covered as orchestration service quality metrics. Figure 5 illustrates that the virtual machine service quality metrics (VM stall, VM premature re
43、lease ratio, VM scheduling latency and VM clock error) apply from “T=0“ when the target VM becomes available to serve VNF users until the VM is released in either an orderly or disorderly way. Figure 5: Virtual Machine Service Quality Metrics ETSI ETSI GS NFV-INF 010 V1.1.1 (2014-12) 135.1 VM Stall
44、VM instances can briefly cease execution, stall or hiccup, for reasons not explicitly requested by the application or cloud consumer, such as “live“ VM migration. During the stall event, the application component hosted by the impacted VM instance does not execute, so both new requests and queued wo
45、rk fails to be served for the duration of the stall event, thereby potentially impacting end user service. Individual VM stall events are measured as the elapsed time between when the VM instance ceases to be executed (e.g. when a VM is paused at the start of a live VM migration) to the instance whe
46、n the VM instance resumes execution (e.g. when VM instance is resumed at the conclusion of a live VM migration event). The primary metrics of VM stall impairments are (1) duration of VM stall event, and (2) frequency of VM stall events. These metrics can be used for, e.g. root cause study as explain
47、ed below in clause 9.2. If execution of a VM instance is suspended for longer than the applications maximum acceptable VM stall time, then the VNFs high availability mechanism will activate to failover user service to another VM instance. As shown in figure 6, stall events that exceed the VNFs maxim
48、um acceptable stall time are counted as premature VM release events rather than VM stall events. 5.2 Premature VM Release Ratio Premature VM release ratio measures the intensity of disorderly VM release events. Premature VM release events explicitly include the following disorderly release event cau
49、ses: 1) NFV infrastructure failure event, like host hardware or hypervisor failure, or power disruption to that host. 2) VM stall events whose duration exceeds the VNFs maximum acceptable VM stall time where the maximum acceptable stall time is an input parameter of the metric. 3) Unavailability or unacceptably poor performance for at least the VNFs maximum acceptable VM stall time for the virtual CPU, virtual NIC or virtual disk resources allocated to the target VM instance. Premature VM release measurements exclude the following relea
