Search

US-12627579-B2 - Availability and failure rate of virtual network functions

US12627579B2US 12627579 B2US12627579 B2US 12627579B2US-12627579-B2

Abstract

There is provided a method, apparatus and non-transitory computer readable media for designing and deploying a network service (NS) meeting availability requirements. The method comprises computing availabilities and failure rates of virtual network functions (VNFs) instances available for deploying the NS. The method comprises designing the NS by defining redundancy and placement of constituents of VNFs instances, using the computed availabilities and failure rates of the VNFs instances. The method comprises deploying the NS.

Inventors

  • Siamak AZADIABAD
  • Maria Toeroe
  • Ferhat Khendek

Assignees

  • TELEFONAKTIEBOLAGET LM ERICSSON (PUBL)

Dates

Publication Date
20260512
Application Date
20220310

Claims (20)

  1. 1 . A method for designing and deploying a network service (NS) meeting availability requirements, comprising: computing availabilities and failure rates of virtual network functions (VNFs) instances available for deploying the NS; designing the NS by defining redundancy and placement of constituents of VNFs instances, using the computed availabilities and failure rates of the VNFs instances; and deploying the NS.
  2. 2 . The method of claim 1 , wherein computing the availabilities and failure rates of the VNFs instances comprises computing, for each VNF instance, the availabilities and failure rates of the VNF instance constituents for each scaling level in accordance with applicable anti-affinity rules and boundary constraints that determine the placement of the constituents of the VNFs instances.
  3. 3 . The method of claim 2 , wherein the applicable anti-affinity rules comprise rules governing sharing of physical resources by defining a distribution of the constituents of the VNFs instances among the physical resources.
  4. 4 . The method of claim 3 , wherein the physical resources comprise physical links, physical hosts and nodes.
  5. 5 . The method of claim 2 , wherein designing the NS further comprises: selecting a minimum availability among the availabilities computed for each scaling level of the VNF instance; and selecting a maximum failure rate among the failure rates computed for each scaling level of the VNF instance; thereby guaranteeing a given availability of the VNF instance for any scaling level.
  6. 6 . The method of claim 5 , wherein selecting a minimum availability among the availabilities computed for each scaling level of the VNF instance comprises: iteratively for each scaling level: computing an availability of each of a plurality of VNF components (VNFCs) for the scaling level; computing an availability of each of a plurality of internal virtual links (IntVLs) for the scaling level; and computing an availability of the VNF instance for the scaling level; and selecting the minimum availability among the availabilities computed for each scaling level.
  7. 7 . The method of claim 6 , wherein the availability, A VNFC , of each of the plurality of VNF components (VNFCs) for the scaling level is computed using: A VNFC =1−(1− A PH *A HV *(1−(1− A VM *A vnfc-App ) k )) n−g*(k−1) *(1− A PH *A HV *(1−(1− A VM *A vnfc-App ) k−1 )) g*k−n where A PH is an availability of a physical host, A HV is an availability of a hypervisor, A VM is an availability of a virtual machine on which the VNFC is instantiated, A vnfc is the availability of a VNFC instance, A vnfc-App is an availability of an application, k is a natural number of VNFC instances that can share a same host, n is a natural number of VNFC instances for the scaling level and g is a natural number of hosts.
  8. 8 . The method of claim 7 , wherein: k = { n , if ⁢ ( no ⁢ anti_affinity ⁢ rule ) ⁢ and ⁢ ( b ≥ n ) b , if ⁢ ( no ⁢ anti_affinity ⁢ rule ) ⁢ and ⁢ ( b < n ) 1 , if ⁢ ( an ⁢ anti_affinity ⁢ rule ) ⁢ and ⁢ ( d ⁢ is ⁢ not ⁢ present ) b , if ⁢ ( an ⁢ anti_affinity ⁢ rule ) ⁢ and ⁢ ( d ≥ b ) ⁢ and ⁢ ( b < n ) n , if ⁢ ( an ⁢ anti_affinity ⁢ rule ) ⁢ and ⁢ ( d ≥ b ) ⁢ and ⁢ ( b ≥ n ) d , if ⁢ ( an ⁢ anti_affinity ⁢ rule ) ⁢ and ⁢ ( d < b ) ⁢ and ⁢ ( d < n ) n , if ⁢ ( an ⁢ anti_affinity ⁢ rule ) ⁢ and ⁢ ( d < b ) ⁢ and ⁢ ( d ≥ n ) where b is a boundary for a natural number of collocated VMs on a host due to capacity limitations of hosts, and d is an allowed degree of host sharing according to anti-affinity rules.
  9. 9 . The method of claim 7 , wherein the availability, A IntVL , of each of the plurality of internal virtual links (IntVLs) for the scaling level is computed using: A IntVL =A intvl if no anti-affinity rule applies; and A IntVL =1−(1− A intvl ) r otherwise; where A intvl is an availability of one instance of IntVL, and r is a total natural number of IntVL instances.
  10. 10 . The method of claim 9 , wherein the availability A vnf of the VNF instance for the scaling level is computed using: A vnf =Π i=1 q A VNFC i *Π j=1 m A IntVL j where q is a total natural number of VNFCs and m is a total natural number of IntVLs.
  11. 11 . The method of claim 5 , wherein selecting a maximum failure rate among the failure rates computed for each scaling level of the VNF instance comprises: iteratively for each scaling level: computing a reliability of each of a plurality of VNF components (VNFCs) for the scaling level; computing a failure rate of each of the plurality of VNFCs for the scaling level; computing a reliability of each of a plurality of internal virtual links (IntVLs) for the scaling level; computing a failure rate of each of the plurality of IntVLs for the scaling level; and computing a failure rate of the VNF instance for the scaling level; and selecting the maximum failure rate among the failure rates computed for each scaling level.
  12. 12 . The method of claim 11 , wherein the reliability, R VNFC , of each of a plurality of VNF components (VNFCs) for the scaling level is computed using: R VNFC = ( R PH * R HV * ∑ i = 1 k ⁢ ( - 1 ) i - 1 * ( k i ) * ( R VM * R App ) i ) n - g * ( k - 1 ) * ( R PH * R HV * ∑ i = 1 k - 1 ⁢ ( - 1 ) i - 1 * ( k - 1 i ) * ( R VM * R App ) i ) g * k - n where R PH is a reliability of a physical host, R HV is a reliability of an hypervisor, R VM is a reliability of a virtual machine on which the VNFC is instantiated, R vnfc is the reliability of a VNFC instance, R App is a reliability of an application, k is a natural number of VNFC instances that can share a same host, n is a natural number of VNFC instances for the scaling level and g is a natural number of hosts.
  13. 13 . The method of claim 12 , wherein the failure rate λ VNFC of each of the plurality of VNFCs for the scaling level is computed using: λ VNFC = - ln ⁢ R VNFC t where t is a time period for which the failure rate is computed.
  14. 14 . The method of claim 13 , wherein the reliability R IntVL of each of a plurality of internal virtual links (IntVLs) for the scaling level is computed using: R IntVL =R intvl if no anti-affinity rule applies; and R IntVL = ∑ i = 1 r ⁢ ( - 1 ) i - 1 * ( r i ) * R intvl i ⁢ otherwise ; where R intvl is a reliability of one instance of IntVL, and r is a total natural number of IntVL instances.
  15. 15 . The method of claim 14 , wherein the failure rate λ IntVL of each of the plurality of IntVLs for the scaling level is computed using: λ IntVL = - ln ⁢ R IntVL t where t is a time period for which the failure rate is computed.
  16. 16 . The method of claim 15 , wherein the failure rate λ vnf of the VNF instance for the scaling level is computed using: λ vnf = ∑ i = 1 q λ VNFC i + ∑ j = 1 m λ IntVL j where q is a total natural number of VNFCs and m is a total natural number of IntVLs.
  17. 17 . An apparatus operative to design and deploy a network service (NS) meeting availability requirements comprising processing circuits and a memory, the memory containing instructions executable by the processing circuits whereby the apparatus is operative to: compute availabilities and failure rates of virtual network functions (VNFs) instances available for deploying the NS; design the NS by defining redundancy and placement of constituents of VNFs instances, using the computed availabilities and failure rates of the VNFs instances; and deploy the NS.
  18. 18 . The apparatus of claim 17 , further operative to compute, for each VNF instance, the availabilities and failure rates of the VNF instance constituents for each scaling level in accordance with applicable anti-affinity rules and boundary constraints that determine the placement of the constituents of the VNFs instances.
  19. 19 . The apparatus of claim 18 , wherein the applicable anti-affinity rules comprise rules governing sharing of physical resources by defining a distribution of the constituents of the VMFs instances among the physical resources.
  20. 20 . The apparatus of claim 19 , wherein the physical resources comprise physical links, physical hosts and nodes.

Description

PRIORITY STATEMENT UNDER 35 U.S.C. S.119(E) & 37 C.F.R. S.1.78 This non-provisional patent application claims priority based upon the prior U.S. provisional patent application entitled “AVAILABILITY AND FAILURE RATE OF VIRTUAL NETWORK FUNCTIONS”, application No. 63/159,224, filed Mar. 10, 2021, in the names of Azadiabad et al. TECHNICAL FIELD The present disclosure relates to the design and deployment of network services. BACKGROUND In the context of Network Function Virtualization (NFV), network functions are provided by Virtual Network Functions (VNF) or by Physical Network Functions (PNF). A Network Service (NS) is the composition of VNFs, PNFs, and/or other NSs that are interconnected by Virtual Links (VL). A VNF is a software implementation of a network function, which is composed of at least one VNF Component (VNFC) and zero, one, or many Internal VLs (IntVL). The VNFCs are the real consumers of the infrastructure resources (e.g., computing and storage resources). A VNF vendor designs and structures the VNF application into VNFCs and specifies the required virtual resources. For each VNF, the vendor creates a VNF Descriptor (VNFD) with its virtual deployment units (VDU). The VNFD also defines VNF Deployment Flavors (VnfDF) to indicate the deployment, connectivity, and virtualized resource requirements of its instances. A VNF is instantiated based on one of its VnfDFs. A VnfDF includes one or many VDU profiles (i.e., VduProfile information element) one for each VNFC, and VL profiles (i.e., VirtualLinkProfile information element) one for each IntVLs. VnfDFs can include affinity and/or anti-affinity rules for the placement of VNFC and IntVL instances. Also, a VnfDF defines scaling levels of the VNF which indicates the number of instances for each VNFC and IntVL at different scaling levels. SUMMARY The availability of a system is defined as the fraction of time the system, a VNF for instance, can deliver its service (or functionality) during a given period. Therefore, the availability of a VNF functionality depends on the outage time of the functionality that the VNF provides. The outage time of the VNF functionality in turn depends on the failure rate of the VNF instance(s) providing it, the failure detection time, and the failure recovery time. There is a need for a method to determine these attributes of the VNF instances as dynamic composite entities. The number of redundant instances for a VNF, if needed, is also determined based on the availability of those instances. Those are needed to be able to design and deploy NSs meeting availability requirements. There is provided a method for designing and deploying a network service (NS) meeting availability requirements. The method comprises computing availabilities and failure rates of virtual network functions (VNFs) instances available for deploying the NS. The method comprises designing the NS by defining redundancy and placement of constituents of VNFs instances, using the computed availabilities and failure rates of the VNFs instances. The method comprises deploying the NS. There is provided an apparatus operative to design and deploy a network service (NS) meeting availability requirements. The apparatus comprises processing circuits and a memory, the memory containing instructions executable by the processing circuits whereby the apparatus is operative to compute availabilities and failure rates of virtual network functions (VNFs) instances available for deploying the NS. The apparatus is operative to design the NS by defining redundancy and placement of constituents of VNFs instances, using the computed availabilities and failure rates of the VNFs instances. The apparatus is operative to deploy the NS. There is provided a non-transitory computer readable media having stored thereon instructions for designing and deploying a network service (NS) meeting availability requirements. The instructions comprise computing availabilities and failure rates of virtual network functions (VNFs) instances available for deploying the NS. The instructions comprise designing the NS by defining redundancy and placement of constituents of VNFs instances, using the computed availabilities and failure rates of the VNFs instances. The instructions comprise deploying the NS. The instructions comprise any of the steps described herein. The methods and apparatus provided herein present improvements to the way design and deployment of network service (NS) operate. BRIEF DESCRIPTION OF THE DRAWINGS FIGS. 1a and b are schematic illustrations of the placement of two instances of a VNFC: a) two VNFC instances of the same profile, sharing the same physical host; b) two VNFC instances of the same profile, using different physical hosts. FIGS. 2a and b are fault tree diagrams (FTDs) of a system with two components: a) the system fails if A and B fail; b) the system fails if A or B fails. FIGS. 3a and b are fault tree diagrams of a system with three components which fails if all