In a general sense, what a service provider wants from an OSS suite hasn’t changed much during the past few years. OSSs are designed to make a service provider’s business and network operate efficiently and competitively. These systems are meant to reduce cost, improve operations and ease service introduction.
Service providers for several years have demanded simplified integration; flexible support for new products, services and processes; customer self-service features; and a superior view of how the business is operating at any given time. With meeting these demands set as a primary development goal, many OSS vendor products have matured to address these requirements, or have come a long way toward doing so.
But with network technology shifting as rapidly as it does, OSS vendors are often criticized for not keeping up with new network technologies. They take heat for being reactive. In fact, a favorite propaganda device for OSS vendors is to accuse each other of failing to support service providers’ business and technical requirements. To quote a somewhat maligned high school football coach, “That’s bulldinky.” Looking at the metro-optical arena for example, not only are OSS vendors delivering effective applications, network equipment providers are incorporating sophisticated OSS functionality into their offerings.
OSS is an imperfect science, plain and simple. OSS vendors all have strengths and weaknesses, but service providers know and account for this. They expect to spend time and money evaluating vendors, and integrating and customizing systems to fit their unique needs. But there are as many deployment schemes as there are service providers, making trends in OSS product uptake rather difficult to evaluate.
A few trends, however, are evident among metro-optical providers in particular. These companies are getting a great deal of attention these days, because they are some of the most successful competitive service providers. They are also deploying what are known as “disruptive” network technologies. A disruptive technology sounds antithetical to a service provider’s goal of delivering reliable service, but what it refers to is any technology that is so effective or promising that it has disrupted network deployment strategies across a range of providers.
Metro-optical providers are built on three of the most significant disruptive technologies we’ve seen in the past decade: Internet Protocol (IP), gigabit Ethernet (GigE) and dense wavelength division multiplexing (DWDM). As such, they have brought both traditional and new demands to the equipment and OSS vendor communities. What’s surprising is how well prepared the market has been to meet those demands. Metro-optical providers are demonstrating some of the most intriguing and advanced OSS capabilities found anywhere, but their approaches to deployment have varied greatly.
Service providers such as Broadwing, Telseon, Yipes!, Cogent and others have deployed a variety of OSS products. These companies’ offerings—as well as their selection of vendors—provide a slightly clearer view of how well the vendors are delivering and where the market for OSS applications could be headed.
Delivering Metro-Optical Service
To understand the role of an OSS, it’s important to understand what goes into delivering a particular service. The architecture of metro-optical gigabit Ethernet services generally is straightforward; multiservice optical nodes on the customer premises connect to optical Ethernet switches to provide access to the DWDM metro backbone. The nodes on the customer premises act as intelligent agents to allow direct control over how bandwidth is allocated to a customer and usually permit dynamic, on-demand bandwidth requests that are filled automatically.
The customer-specific provisioning intelligence resides in the access device. The equipment-specific or domain provisioning system communicates with the access device and the serving switch to configure a connection and allocate bandwidth, often in response to requests coming from a customer self-service Web portal. The equipment-specific or domain provisioning system is generally a vendor-provided OSS.
As mentioned earlier, not only have OSS vendors responded to the call for applications and simplified integration, but some equipment vendors have advanced their own offerings, as in the case of Broadwing’s deployment.
Making Broadwing Tick
From the outside, and according to company announcements, Broadwing appears to have installed a multitiered OSS architecture with provisioning functions for metro-optical IP VPN services divided between two key systems. Broadwing announced that it implemented CIENA’s CoreDirector product suite, which includes optical access devices and metro switches, as well as a significant set of bundled OSS applications for service creation, provisioning and integration. CIENA’s provisioning application in this instance, therefore, is the optical Ethernet domain manager.
Broadwing has also implemented Ai Metrix NeuralStar, an optical-friendly, IP-world application for provisioning and network management. Keep in mind that Broadwing offers a range of services, including SONET connectivity, voice services, traditional T-1s, ATM and frame relay. NeuralStar is a framework with components designed to support inventory, provisioning and network management functions for all such data and optical networks. The application boasts the capability to support products from about 18 different equipment vendors, 16 of which you could probably name off the top of your head. Not surprisingly, CIENA is on that list.
The combination of technologies from Ai Metrix and CIENA enables Broadwing’s IP VPN customers to reserve and provision bandwidth connections, or sessions, in near-real time through a Web interface, and also provides performance management functions—likely derived from CIENA’s service layer manager application.
The Requirements Report Card
Based on the fundamental requirements common to most service providers, Broadwing’s vendors seemed to have delivered quite adequately.
First, ease of integration—a no-brainer. CIENA offers an integration server and APIs for provisioning, fault management, inventory and VPN configuration. More importantly, Ai Metrix and CIENA have an established relationship to integrate their products. In the future Broadwing shouldn’t have a problem, because a customer of such scale can dictate requirements to some extent, and will benefit from the vendors coordinating their releases. Keep in mind that all of this technology isn’t necessarily simple to work with, but that’s also not the expectation. We are in fact talking about some of the most complex software and network systems in the world. The important point is that we’re seeing vendor collaboration and a significant effort to deliver integrated, highly functional systems.
Another key requirement is support for a multitechnology network. In this case, CIENA doesn’t really apply, because it represents its own technology domain. In addition, Broadwing isn’t just some startup, it’s really an ILEC—Cincinnati Bell. While it has grown far past its roots, it still maintains a legacy network base that offers ATM, frame relay, GigE and SONET services. Ai Metrix passed the multitechnology test just by Broadwing’s willingness to license its product. The company, however, has received consistent accolades for its ability to manage and provision multiple network layers and technology sets.
Finally comes the kicker: the ability to support customer self-provisioning and service level agreements. This kind of requirement reflects an overall desire by service providers to reduce cost and activation time while promoting customer loyalty. CIENA’s systems, as discussed, enable Broadwing to let customers dynamically configure IP VPN connections and also view associated performance data. Broadwing can also deliver that information to upstream systems that can react to various SLA status conditions. This is a value-added service Broadwing can offer to its customers that is entirely based on the capabilities of its OSSs. In other words, the vendors are doing their jobs.
Observations
What’s surprising is that the equipment vendor, CIENA, has provided a fully featured OSS package that’s optimized for its network product and is integration-friendly. It wasn’t too long ago that equipment vendors ignored OSS and offered little more than SNMP management interfaces to their gear. Now CIENA has sold its CoreDirector provisioning package to several operators, including Broadwing, international optical provider Tycom and Flag Telecom.
But the OSS and equipment vendors are really meeting in the middle. Equipment vendors are creeping closer to the back office while OSS vendors get closer to the network. This is beginning to culminate in a distinct trend toward service resource-focused systems that maintain a holistic view of a provider’s network, customers, services and inventory. Equipment vendors tend to draw the line at business intelligence, sticking to the functions necessary to operate their gear. OSS vendors have generally specialized in automating business processes, but as automated network design and configuration become more prevalent, the need grows for synchronizing business processes and network operations.
Yipes! has already received a great deal of press for its OSS efforts, which entail more traditional best-of-breed selection. The company chose OSS vendors including Syndesis and Eftia to provide its service management backbone. Cogent employs a number of OSS capabilities from its equipment vendors, including Cisco Systems, Extreme Networks and LuxN.
Yipes! doesn’t represent Syndesis’ only win in the optical IP arena. Syndesis has an OEM deal with Cisco and serves as the provisioning application for Cisco’s ONS 15454 optical transport and related systems, which happen to provide part of the core of Cogent’s optical DWDM network. Cogent has also deployed gigabit Ethernet technology from Extreme Networks, another vendor Syndesis supports, running over DWDM gear from LuxN.
LuxN provides its own optical layer provisioning and SLA configuration and management platform called Optical Service Level Management (OSLM). This enables on-demand provisioning of DWDM services, enforcement of SLA measurements and policies at the optical layer, and support for multiple Layer 2 transport technologies such as GigE or TDM-based OC-x transmissions. OSLM also supports CORBA and XML interfaces for integration to other OSSs. Once again we have an example where the equipment manufacturers are delivering strong OSS functionality that can be exploited from higher level applications. Syndesis is unusual in that its platform can be both a stand-alone activation application and an equipment-specific domain manager.
Syndesis has proven its ability to support multitechnology networks, and though it is especially strong for Cisco gear, Syndesis also maintains relationships with Lucent (with whom it also has an OEM arrangement), Nortel and DWDM leader Sycamore Networks. Syndesis is also particularly strong in supporting access network technologies, notably DSL and multiservice edge devices. From the service provider’s perspective, Syndesis is a safe bet, and its relationships with companies like Cisco, Lucent and Extreme Networks have established it among the leaders in the service activation space.
An Eftia Surprise
Landing Yipes! is a huge win for Eftia, a company many had written off as being on its way out. The company has been pretty well funded but also has taken a good share of punishment in the marketplace. The philosophy Eftia has touted for some time could be proving itself, though, and a number of competitors have followed.
Eftia has focused on data and IP since its inception several years ago. The knock on its application was never a lack of functionality but rather difficulty of configuration and implementation. That difficulty was reportedly tied to the customer’s need to perform a detailed network audit in order to build the strong inventory store the application would need to work its magic. Yipes!, a growing startup, could build its inventory database as it builds out its network. It was not shy in announcing its selection of Eftia to provide service management for its IP-over-optical offerings. Also, Eftia partners with Syndesis, bringing some preintegration to the table and thus meeting a primary service provider requirement.
As mentioned, seeing Eftia win this deal bolsters an emerging trend away from traditional systems for inventory, network design and provisioning to a better integrated service resource store with physical and logical inventory, customer specific views of network services, assign-and-design functionality, and even fault management.
In addition, Cogent’s selection of Visionael, also an open inventory-based service management OSS, further reinforces the shift. The service resource store, depending on the vendor, is generally an open inventory database that is configurable to support any kind of network resource throughout the entire network hierarchy.
A Quick Look At Telseon
It’s difficult to get a clear view into Telseon’s back office, but it appears the company is leaning somewhat on an integrated hardware and OSS package from ONI Systems. Telseon and Dynegy are in the process of building out an 18-city all-optical network on ONI’s DWDM transmission equipment. Telseon extends Web-based self-provisioning of DWDM virtual private optical networks (VPONs) to its customers, apparently based on ONI’s OPTX management portal. The portal is part of a suite that goes beyond activation and SLA management for optical VPNs. OPTX also includes network design tools, a Java-based element management platform, a CORBA-based integration gateway and an optical link manager that optimizes interaction between the data and optical transport layers.
Telseon has gone so far as to brand its Web-based provisioning service as the Telseon IP Provisioning system. Though the company hasn’t revealed how exactly it handles provisioning at the IP layer, it did announce that it selected Visionael to provide its network inventory management suite, marking yet another win for service resource-based systems. Telseon describes its provisioning system in a white paper, and it seems the company gives its customers direct access to a network design wizard that enables graphical configuration of VPON pathways, including class of service features, bandwidth profiles and customizable naming conventions for logical paths.
This example highlights the growing movement, particularly in the area of disruptive technologies, of equipment vendors to build sophisticated, domain-specific OSSs that will co-exist in an integrated OSS environment. Such applications are often tied directly back to a network configuration or service resource store. The intent is to have little or no delay between designing and assigning a service, activating it on the network, and updating the network inventory store to reflect the changes in the network’s and customer’s states. Again, the service resource store provides the core for the entire OSS suite and is the place where customers, technologies and services are all related.
Service Resources
Leo Szilard, a leading 20th century physicist credited with discovering the chain reaction, was recognized for his enlightening scientific philosophy. One of his personal Ten Commandments was to look at the connections among things in order to understand the purpose of each element in an overall system, and thus better understand the entire system itself. This can apply to almost any system in the universe, but it’s especially true for the direction provisioning is taking.
Provisioning has taken many forms over the years, but from a traditional point of view it’s often considered the processes through which a service provider designs and assigns a service in some kind of network configuration store. It has grown beyond that to include management of all of the tasks required to fulfill a service order, including functions such as establishing SLAs and activating billing.
The problem with traditional provisioning systems, however, is that they were often little more than standalone design tools, or were actually collections of weakly integrated or disparate systems, each fulfilling a specific role for a different network layer or service technology. As a result, providers could not easily examine the relationships between customers, services, network performance and available resources. This lack of integration generally leads to stranded equipment, unbilled services, reactive customer care and other business inefficiencies.
Several years ago, forward-looking vendors such as Granite Systems began altering perceptions of provisioning and network inventory. Provisioning and inventory are now collectively about service resources and understanding the connections among all the elements in a provider’s end-to-end service offering. Provisioning tasks are now performed directly with the network resource store, averting discrepancies among various network databases. This combined view allows the service provider to derive intelligence about its operations that was previously impossible.
Dynamic service provisioning is slowly becoming the norm, especially as on-demand, multiservice, CPE-based packet and wavelength offerings grow prevalent. So it makes sense to move to a service resource model, since more of the network can be provisioned automatically and remotely than was the case in the circuit world. In the old days, everything from setting up a new dedicated circuit to ordering a cross-connect required some kind of manual engineering design task followed up by a manual equipment installation, cross-connect patch, loop connection, and so on. Provisioning times were, and generally still are, measured in days, weeks or months, because operational efficiency just didn’t matter as much. Today, efficient service delivery is a priority—moreover, a necessity—for service providers, making the service resource concept very attractive.
