Patent Application
20050071453
The invention relates generally to communications systems. More particularly, the invention relates to a system and method for performing service performance correlation (SPC) and service fault correlation (SFC) to manage services transported across a communications network in support of a service level agreement (SLA).
Transport networks of today need to provide cost-effective transport for various types of client information, including multi-service traffic ranging from synchronous traffic (e.g., DS-1, DS-3, and STS-12) to asynchronous traffic (e.g., Internet Protocol (IP), Ethernet, and Asynchronous Transport Mode (ATM)). Increasingly, service providers are operating their circuit-oriented services over transport networks based on synchronous optical network (SONET) or synchronous digital hierarchy (SDH) and their connectionless services over packet transport networks based on Ethernet, Multi-Protocol Label Switching (MPLS), IP, or combinations thereof.
In the offering of such services, service providers generally operate large networks supporting numerous services over a basic network infrastructure comprised of various network layers. These layers include physical plant (i.e., conduit and fiber), layer 0 optical links and amplifiers, layer 1 SONET/SDH synchronous optical networking, and layer 1 asynchronous digital networking or digital hierarchy (DS0, DS1, DS3). Layer 2 networks include X.25, Frame Relay, ATM, and MPLS. Layer 3 networks include both private IP and public Internet. When offering services across these networks it is often difficult to correlate network faults with service faults. To effectively manage their customer relationships and offer effective services, service providers want to be able to monitor their services to ensure that each service is performing in accordance with its corresponding service level agreements (SLA). As part of ensuring conformance to an SLA, service providers need to ensure Service Fault Correlation (SFC) and Service Performance Correlation (SPC). Service Fault Correlation entails correlating a failure or fault in the service provider's network with each affected service by the failure or fault. Service Performance Correlation entails correlating the performance of network traffic with a particular service.
Traditionally, responsibility for the proper operation of the transport network resides with a network operational support system (OSS), also referred to as network management. Network management performs a variety of management functions, including fault management, configuration or provisioning management, accounting, performance monitoring, and security. To accomplish these functions, the network elements in the transport network collect or generate information to be made available to the network management. This information is indicative of the functional performance of the transport facility (i.e., the network elements, paths, and links in the transport network), and is referred to as network-related information.
In contrast to the roles of the network management, service providers are responsible for order fulfillment, service assurance, and billing. In effect, the service providers manage the customers of their services and maintain customer relationships; when customers are experiencing problems with their services, they interact with their service providers. Often problems occur in the transport network of which service providers are unaware, unless notified thereof by a customer or alerted thereto by the network management.
To answer a customer's inquiry regarding a service, the service provider would ideally be able to confirm the problem directly. Nonetheless, a service provider typically needs to consult with network management to obtain information necessary to corroborate or refute the customer's problem. Even then, the network-related information obtained from the network management does not directly identify any specific service. Consequently, the service provider must refer to other sources of information, such as telephone logs of customer calls and databases cross-referencing network resources to services, to piece together the full picture of a service's performance. Service providers encounter this same difficulty when network management alerts them of problems encountered in the transport network; the network-related information given to the service providers does not directly identify any affected service.
This inability to ascertain their services' performances handicaps service providers in the execution of their other service management functions. Without the ability to monitor a service's performance, it is difficult to determine if the service is performing in accordance with the SLA. Consequently, current SLAs tend not to specify significant detail about the service. Similarly, service providers-lack service-specific metrics by which they can design (i.e., engineer) their transport networks for supporting the services properly. Moreover, the difficulty in correlating network problems to specific services complicates the task of billing customers accurately for their purchased services; customers expect to pay less than full price for services that did not perform as expected, but service providers experience inefficiencies when attempting to verify service downgrades and outages.
Traditionally, the correlation of faults (SFC) and service performance (SPC) have been relegated to the OSS support systems and to the manual correlation of records. This process results in delayed understanding of service performance and SLA conformance. Thus, there is a need for a system and method that enable service and network management functions to be more effectively executed in the network rather than current techniques that occur at higher layers above the network.
In one aspect, the invention features a method for managing a service transported over a transport network. Service traffic from a client network is received at a first network element. A performance of the service traffic is measured at the first network element. The service traffic is transmitted from the first network element to a second network element. A performance of the service traffic received from the first network element is measured at the second network element. The performance of the service traffic measured at the first network element is correlated with the performance of the service traffic measured at the second network element to provide an end-to-end performance correlation of the service across the transport network.
In another aspect, the invention features a method for managing a service transported over a transport network. Traffic associated with at least one service is transmitted over the transport network. A network element detects a condition occurring in the transport network and correlates a facility element of the transport network and at least one service affected by the condition.
In yet another aspect, the invention features a communications network comprising a first network element connected to a transport facility. The first network element receives service traffic from a client network, measures a first performance of the service traffic, and correlates the measured first performance and a service to produce a first performance of service (PoS) service metric. A second network element is in communication with the first network element over the transport facility to receive service traffic transmitted over the transport facility from the first network element. The second network element measures a second performance of the service traffic and correlates the second measured performance and the service to produce a second PoS service metric. The network also includes a service management channel between the network elements over which one of the network elements transmits the PoS service metric measured by that network element to the other of the network elements.
The above and further advantages of this invention may be better understood by referring to the following description in conjunction with the accompanying drawings, in which like numerals indicate like structural elements and features in various figures. The drawings are not necessarily to scale, emphasis instead being placed upon illustrating the principles of the invention.
Networks constructed in accordance with the present invention provide service providers with a variety of mechanisms for more closely managing the services purchased by their customers. A service, as used herein, is a guarantee of transport of customer-offered traffic with specific performance commitments. The service provider and possibly one or more carriers transport the customer-offered traffic over a transport network between service-termination points. Network elements at these service-termination points measure the performance of the service based on characteristics of the customer-offered traffic and correlate the measured performance with a specific service, to produce a performance of service metric (hereafter, referred to generally as PoS). In some embodiments of the invention, the service termination points collect and exchange PoS service metrics across the network over a service management channel. Having service metrics from both service termination points enables service providers to compare performance at both ends of the service.
From PoS metrics, service metrics called availability of service (AoS) and rate of service (RoS) are computed. Using these PoS, AoS, and RoS service metrics, service providers have additional tools for customizing service level agreements (SLAs) and service level specifications (SLSs) for their services, for pricing their offered services, and for determining whether their services are performing in accordance with the SLA and SLS.
The invention also enables network failures and faults to be correlated with specific services either in service reports that correlate service performance with the transport facility or in network reports that correlate performance of the transport facility with services. Personnel can obtain service metrics, service reports, and network reports directly from a network element at a service termination point or at an interior point in the network. This information is available in real-time; service personnel can obtain the information themselves directly and immediately respond to customers' service inquiries, without having to take the traditional routes involving network management.
The LAN extension network 12 can be a connectionless or connection-oriented local area network (LAN), metro-area network (MAN) or wide-area network (WAN) or any combination of connectionless and connection-oriented networks. An example of a connectionless network is the Internet. Communication over a connectionless network occurs according to the particular technology employed, examples of which include, but are not limited to, Ethernet, MPLS, IP, and combinations thereof. For example, MPLS can provide connection-oriented processing above a connectionless IP network. An example of a connection-oriented network is an optical network based on a synchronous data transmission standard such as Synchronous Optical Network (SONET), Synchronous Digital Hierarchy (SDH), or Optical Transport Network (OTN)). The LAN extension network 12 can itself be a network of networks, spanning a plurality of different service providers and carriers and including both connectionless and connection-oriented sub-networks.
A near-end edge service network element NE 14 is in communication with a far-end edge-service network element NE 16 through an intermediate or core network element NE 18. Communication among the NEs 14, 16, 18 is over a transport facility 20, in accordance with the particular standard used for data transmission: for example, SONET for an optical transport facility, IP for a packet-switched transport network. Transmission of Ethernet services over an optical transport facility, for example, is referred to as Optical Ethernet. Other types of transport facilities than optical transport can be used, such as wired or wireless transport, without departing from the principles of the invention.
Each NE 14, 16 is also in communication with a respective interface (not shown) at a demarcation point for communicating with one of the LANs 10; NE 14 communicates with the LAN 10-1, and NE 16 communicates with LAN 10-2. A service provider or carrier uses the LAN extension network 12 to support a service purchased by a customer under terms governed by an SLA. The NEs 14, 16, and 18 illustrate an oversimplified path by which traffic for a particular service traverses the LAN extension network 12. For optical transport networks, the NEs 14, 16 operate as add/drop multiplexers (ADMs) and customer traffic pertaining to the service travels from one edge-service NE 14 to the other edge-service NE 16 through the core NE 18 over a dedicated circuit or path. For connectionless networks, the NEs 14, 16 are, for example, Ethernet edge switches, and the service traffic is routed from the NE 14 to the NE 16 through the core NE 18. More than one service can traverse this particular set of NEs 14, 16, 18, although to simplify the description of the invention only a single service is described. The demarcation points are the termination end-points of the service in the LAN extension network 12. In one embodiment, each NE 14, 16 comprises the respective interface at the demarcation point. Other NEs in the LAN extension network 12 for carrying traffic of other services are not shown.
The near-end NE 14 associates a unique service identifier (hereafter, service ID) with the service traffic. The association of the service ID to the service traffic can occur at layer 1 (e.g., SONET) or at layer 2 (e.g., packet/Ethernet). An example of a layer 1 implementation is to maintain a table or database that cross-references service traffic arriving at a particular port of the near-end NE 14 with the service ID. An example of a layer 2 implementation is to include the service ID in each packet of the service traffic. The service management 30 or network management 28 can install the service ID at the near-end NE 14 when the service is provisioned.
Each edge-service NE 14, 16 computes counts for various performance parameters based on characteristics of the service traffic. In one embodiment, the particular counted performance parameters are taken from a standard set of Management Interface Base II (MIB II) objects (e.g., INFRAMES) defined for managing TCP/IP networks. Traditionally, these performance parameters serve as a measure of performance at a port (e.g., Ethernet or Fiber Channel port) in the client network 10-1, but are not correlated to any service in particular. In accordance with the principles of the invention, service providers correlate these counts (i.e., performance metrics) to a particular service. More specifically, the NE 14 uses the service ID to correlate computed performance metrics with the specific service from which the counts were obtained. This correlation of a performance metric to a specific service produces a performance of service (PoS) attribute or service metric that can be monitored by the service provider. Examples of PoS service metrics are given in more detail below.
When service providers derive PoS service metrics from standard performance metrics, such as the MIB II objects, customers and service providers can compare their independently computed service metrics to verify each other's evaluation of the service performance. For instance, a customer can measure the performance of a purchased service within the client network 10-1 based on counts obtained for certain MIB objects obtained at an Ethernet port. The service provider, based on the same MIB objects, measures the performance of the service within its own LAN extension network 12. With its own gathered performance metrics, the customer can verify service performance reports from the service provider, and the service provider, with its own gathered service metrics, can confirm or deny service inquires raised by the customer. Moreover, as described below, other service metrics derive from PoS service metrics, thus providing additional service metrics that can be computed and similarly compared by customer and service provider.
In one embodiment, some service-related communication between the edge-service NEs 14, 16 occurs over a service management channel (SMC) 22. For circuit-oriented communications over an optical network, exemplary implementations of the SMC 22 include using 1) path overhead (POH) bytes, and 2) Client Management Frames (CMF) of the Generic Framing Procedure (GFP). These implementations of the SMC 22 are described in U.S. patent application Ser. No. 10/666,372, filed Sep. 19, 2003, titled “System and Method for Managing an Optical Networking Service,” the entirety of which patent application is incorporated by reference herein. For connectionless communications, an exemplary implementation of the SMC 22 is with Ethernet Virtual Connection (EVC) technology. In general, an EVC is an association between two or more User Network Interfaces (UNIs), where a UNI is a standard Ethernet interface at the demarcation point. Each EVC limits the exchange of Ethernet packets between those UNIs that are part of the same EVC.
Over either type of SMC 22 (i.e., circuit-oriented or connectionless), the NEs 14, 16 periodically exchange performance monitor (PM) reports and service metrics, such as the PoS, RoS, and AoS service metrics described herein. Further, for embodiments having the SMC 22, the core NE 18 can be configured to access and process the contents of the SMC 22 (the core NE is referred to as being service-aware). Accordingly, the core NE 18 serves as a portal for monitoring the service between the service termination points. An operator of the network management 28 or service management 30 can use the core NE 18 to intercept the service PM reports and service metrics produced by the edge-service NEs 14, 16 and collect historical performance information (i.e., service monitoring). The near-end NE 14 and the core NE 18 can accumulate and store service-related information for a predetermined length of time, and with the accumulated information, the operator can evaluate the performance of the service against an SLA. As described herein, service-related information includes the service state of the interfaces with the client networks 10, the path state, PM reports, service reports that correlate service degradation with the transport facility, and network reports that correlate network faults to affected services.
To enable access by the network management 28 and service management 30 to the service-related information, a computing system 24 is in communication with the near-end NE 14 or, alternatively, with the core NE 18. Network management 28 is also in communication with the service management 30. As described above, the network management 28 is responsible for performing fault management, configuration or provisioning management, accounting, performance monitoring, and security (i.e., FCAPS) for the LAN extension network 12. The service management 30 is responsible for the fulfillment, assurance, and billing (FAB).
In one embodiment, the LAN extension network 12 is a circuit-oriented optical network based on an optical transmission standard, such as SONET. In this embodiment, the encapsulation unit 66 applies Generic Framing Procedure (GFP) and virtual concatenation (VCAT) technologies to the service traffic, and the adapter 68 multiplexes the service traffic on an appropriate Synchronous Transport Signal (STS) path. Service messaging between the NEs 14, 18 occurs over the SMC 22 embodied, for example, by the GFP/CMF or by the POH.
In another embodiment, the LAN extension network 12 is a connectionless network based on packet-based transport (e.g., Optical Ethernet and MPLS). Ethernet Virtual Connections (EVCs) or MPLS Label Switched Paths (MPLS LSPs) can operate to implement the SMC 22. The encapsulation unit 66 provides layer 2 encapsulation using one of a variety of encapsulation mechanisms such as, for example, Media Access Control (MAC) in MAC (MiM), MAC in MPLS, MAC in Point-to-Point Protocol (PPP), with or without Q-tag insertion. The network adapter 68 transmits the service traffic over an appropriate EVC or MPLS LSP.
The NEs 14, 16 also include a processor 70 for measuring performance metrics, computing PoS service metrics from the performance metrics (examples listed below), and computing RoS and AoS service metrics from the PoS service metrics, as described in connection with
Performance metrics measured by the near-end NE 14 are of two types: ingress and egress performance metrics 50, 54, respectively. Ingress performance metrics 50 are collected from service traffic signals entering into the interface 64 and into the encapsulation unit 66. Examples of ingress PoS service metrics for Ethernet Private Line (EPL) and Fiber Channel services, computed from performance metrics measured from service traffic signals entering the interface 64 and correlated to a specific service, are listed in Table 1 below.
Examples of ingress PoS service metrics for Ethernet Private Line (EPL) and Fiber Channel services, computed from performance metrics measured from service traffic signals entering the encapsulation unit 66 and correlated to a specific service, are listed in Table 2 below:
Egress performance metrics 54 are collected from service traffic signals exiting from the interface 64 and from the encapsulation unit 66. Examples of egress PoS service metrics 54 for Ethernet Private Line (EPL) and Fiber Channel services corresponding to service traffic in the interface 64 are shown in Table 2 above and examples of egress PoS service metrics corresponding to service traffic in the encapsulation unit 66 are shown in Table 1 above. Other PoS service metrics can be collected in addition to or instead of those listed above without departing from the principles of the invention.
The near-end NE 14 also detects ingress and egress alarms 58, 62, respectively, and correlates such alarms to a specific service. Ingress alarms are obtained from service traffic signals entering the interface 64. Examples of ingress alarms for an Ethernet Private Line (EPL) service are shown in Table 3 below, including an example of their associated severity and an associate responsive action (taken by the NE 14):
Egress alarms are obtained from service traffic signals exiting from the interface 64, from the encapsulation unit 66, and from the network adapter 68. Examples of type ingress alarms at the interface 64 for an EPL service are shown in Table 4 below, including an example of their associated severity and an associated responsive action taken by the NE 14:
Table 5 below lists examples of egress alarms detected at the encapsulation unit 66, including their associated level of severity and an associated responsive action taken by the NE 14:
Table 6 below lists examples of egress alarms detected at the network adapter 68, including their associated level of severity and an associated responsive action by the NE 14:
The far-end NE 16 computes the same types of ingress and egress performance metrics 50′, 54′ and monitors for the same types of ingress and egress alarms 58′, 62′ as the near-end edge service NE 14.
The RoS service metric provides a mechanism by which the service provider can measure the actual bandwidth utilization of the service. In general, each RoS service metric 76 is a PoS service metric measured over a specific time interval. For example, if the PoS service metric is INOCTETS, the corresponding RoS service metric is, for example, INOCTETS per second. As another example for measuring RoS, if the PoS service metric is INFRAMES, the corresponding RoS service metric is, for example, INFRAMES per second. Further, the RoS service metric provides a measurable basis by which a cost can be associated with the service. Accordingly, each RoS service metric is a billing attribute that can be specified in both the SLA and SLS.
For Ethernet (and WAN) services, the AoS service metric 78 is determined from the INFRAMESERR and INFRAMES parameters; for Fiber Channel services, the AoS service metric is based on the INOCTETSERR parameter. In brief, a service is deemed to have become unavailable (UnAvailable Service or UAS) upon the occurrence of ten consecutive severe errored seconds (SES). An SES is deemed to have occurred for an Ethernet or WAN service when the percentage of incoming frames for a one-second interval exceeds 1%. For Fiber Channel, more than 500 octets having errors in a one-second interval causes an SES. Table 7 below shows summarizes the conditions defining AoS for Ethernet, WAN, and Fiber Channel services.
In embodiments having the SMC 22, the far-end NE 16 periodically transmits its computed PoS, RoS 76′, and AoS 78′ service metrics to the near-end NE 14. Having the service metrics from both ends of the service, the processor 70 of the NE 14 can compare near-end RoS with far-end RoS and near-end PoS with far-end PoS. To determine an overall AoS applicable to the full span of the service (in contrast to the AoS computed for a single node, described above), the NE 14 combines the AoS statuses of both NEs 14, 16. For example, if the AoS service metric is UAS (i.e., unavailable) at either or both of the NEs 14, 16, the overall AoS for the service is unavailable.
Having access to service metrics gathered at both ends of the service, service providers can now specify and verify quality of service (QoS) and class of service (CoS) service metrics in their SLAs and SLSs that account for service performance at both service ends (i.e., not only at egress at the near-end NE 14, but also at ingress at the far-end NE 16). For example, a service provider can now specify different end-to-end RoS service metrics to define different levels of CoS: that is, services guaranteed a higher RoS are deemed to have a higher CoS. Similarly, different end-to-end AoS service metrics can serve to define different levels of QoS.
Similar to the service messaging from the far-end NE 16 to the near-end NE 14, the near-end NE 14 transmits its computed service metrics to the far-end NE 16. The core NE 18, if service-aware, can intercept service messages traveling in both directions, and thus perform similar end-to-end comparisons between RoS and PoS service metrics and end-to-end evaluations of AoS, CoS, and QoS as those performed by the near-end NE 14. Accordingly, network management 28 and service management 30 are each able to access the core NE 18 to review this information.
The overall nodal service state 80 depends upon the individual states of the client link, user network interface (UNI), path, and far-end client signal. For the overall service state to be “green” (i.e., operating in accordance with the SLA) the individual states need to be “OK” for each of the client link, user network interface (UNI), path, and far-end client signal. If any of the listed facility elements fail, the overall service state 80 becomes “red.” If any of the link, UNI, or path degrades, the service state becomes “yellow.” When the overall service state 80 is “red,” the NE issues a network alarm report that correlates the failing facility element with the affected service. For example, if the NE 14 determines that the path has failed, the NE 14 issues an alarm report identifying the facility elements on the failed path and each service affected by the failure. Consider for example that the path passing from NE 14 to NE 16 through NE 18 has failed, the network alarm report may read “Path 14-16-18 is failed, affecting services with service identifiers GR100 and NR104.” The network alarm report is made available to the network and service management 28, 30 (by way of the computing system 24). With this information, the network management can readily determine the failing network resources, and the service provider can readily correlate the network problem to specific affected services. Note, for purposes of this description, network faults are considered included in the term network failure, although network faults and network failures differ at least in their degree of severity.
The NE 14 also generates a service alarm report if the service degrades. For example, for a path that degrades, the service alarm report may read “Service with service identifier GR110 is degraded on the path through network elements 14, 18, and 16.” Service alarm reports, like network alarm reports, are made available to the network and service management 28, 30.
In embodiments having the SMC 22, the far-end NE 16 periodically transmits its individual service states to the near-end NE 14 (or its overall nodal service state). The near-end NE 14 combines its individual states with those of the far-end to produce an overall end-to-end service state 80′. Table 9 below shows the end-to-end service state 80′ as determined by the state information obtained from both the near-end and far-end NEs 14, 16.
The end-to-end service state 80′ depends upon the individual states of the client link, user network interface (UNI), path, and far-end client signal for both the near-end and far-end NEs 14, 16. For the end-to-end service state 80′ to be “green,” every one of these individual states need to be “OK”. If any of the listed individual states fail, the end-to-end service state 80′ becomes “red;” if any degrades, the end-to-end service state 80′ becomes “yellow.” Network and service alarm reports of network failures and degraded services are like those described above for a single edge service state.
The near-end NE 14 computes (step 108) performance metrics and correlates (step 112) these performance metrics with a specific service. Using the computed performance metrics, the near-end NE 14 computes (step 116) various service metrics of the invention.
In an embodiment having the SMC 22, the near-end NE 14 sends (step 118) its computed service metrics to the far-end NE 16 and receives (step 120) from the far-end NE 16 service metrics computed by the far-end NE 16. For this embodiment, the near-end NE 14 can compare the service metrics determined at the near-end with those determined at the far-end to obtain an end-to-end account of the performance of the service. For example, if the near-end NE 14 counted 1000 OUTOCTETS transmitted for a particular service, but the far-end NE 16 counted only 990 INOCTETS received, then the performance of the end-to-end service is seen to be 99%.
In step 124, the service metrics are reported to the service management 30. Such reporting of service metrics are referred to as service performance correlation (SPC) reports. The reporting can be active or passive; by active meaning that the near-end NE 14 automatically transmits a report to the computer system 24, by passive meaning that the service metric information is stored (for a predetermined interval) at the NE 14 until service management 30, for example, accesses the NE 14. The SPC report can be a nodal report based on those service metrics as determined by the NE 14 only, or an end-to-end report based on service metrics as determined at both ends of the service. With these reports, the service management 30 can compare (step 128) the current performance of the service with the values used in the SLA. The service management 30 can also compare (step 132) the current performance of the service with the values defined in the SLS.
While the invention has been shown and described with reference to specific preferred embodiments, it should be understood by those skilled in the art that various changes in form and detail may be made therein without departing from the spirit and scope of the invention as defined by the following claims.
This application claims the benefit of the filing date of U.S. Provisional Application Ser. No. 60/507,278, filed Sep. 30, 2003, titled “Structured Addressing for Optical Service and Network Management Objects,” the entirety of which provisional application is incorporated by reference herein.
| Number | Date | Country | |
|---|---|---|---|
| 60507278 | Sep 2003 | US |
