Most WiMAX networks operating today have, at best, rudimentary OSS (Operations Support Systems).  This puts them at a clear disadvantage in respect to established landline networks, almost all of which utilize extensive suites of OSS software.

Everyone knows what OSS is-sort of.  Few, however, including many network engineers, understand the intricacies of today's many and varied product offerings, or how OSS impacts the bottom line and indeed the network's very survival.  And, according to many of those interviewed for this article, no one is more deficient in regard to OSS expertise than the WiMAX operator, especially the independent.

Needless to say, we can't explore all of those intricacies in the course of a single article, but we can enumerate the basic categories of software, explain their functions, and illustrate their benefits.  And those benefits, as we shall see, are considerable. 

It is not the intent of this article to recommend the products of specific OSS software companies, but rather to discuss the form of a well integrated system, and to provide a basis for evaluating distinct software offerings.  OSS at present is a very volatile industry, and one marked by waves of consolidation, and today's software products may not be tomorrow's.  But the basic needs of the network operator are unlikely to change substantially in the next year or in the next two years,

OSS - a Basic Definition

OSS, by the broadest definition, consists of network management software, that is, software that assumes many of the routine administrative functions traditionally performed by technicians and clerical staff.  Such functions include discovery of network resources for the purpose of providing bandwidth to meet customer demands;  network element configuration;  service provisioning and activation;  customer authentication;  providing assurance-that is, determining that desired changes in the network have in fact been implemented;  enforcement of service level agreements;  and, by some definitions, billing.  Most definitions would also include a more nebulous functionality called mediation which has to do with integrating subsystems into a synergistic whole, so that, for example, changes in a customer's service profile are immediately reflected in his or her billing information and in the allocation of network resources to serve that individual.

The Current State of OSS and What It Means for the WiMAX Operator

OSS must be counted as a fairly recent innovation in the telecom space.  Prior to the nineteen nineties, most networks, whether landline, mobile, or cable, were not automated to any significant extent.  Provisioning and changes in services were accomplished by means of manual interventions involving customer data bases, while network elements were manually configured to reflect those changes.

This state of affairs reflected both the limited range of service offerings available, and the de facto monopoly that prevailed on the local level.  If no one else was offering local phone service or video programming, then the network operator had little incentive to make the operation more efficient.  Any costs associated with inefficiencies of manual functions could simply be passed on to the subscriber.  And the fact that so few service options were presented to the customer meant that service provider was freed from the requirement of recalibrating the network to meet individual needs each time a subscriber requested a modification of his or her service bundle.  Obviously, basic access with no guarantees as to service quality places few demands on either the network or the network manager.

As variegated business services such as conferencing, whiteboarding, and VPNs became the norm, and, sometime later, triple and quadruple plays appeared in residential markets, the burden of network management grew apace.  Not only were competing demands being made upon available bandwidth, each individual subscriber terminal now had to be enabled to support any of a multitude of different service packages.  And the situation was made worse by the fact that the devices permitting new services within the network might very well emanate from different vendors than the devices already anchoring the core of the network and must therefore be coordinated through the agency of additional middleware.  Rather suddenly, the need for OSS products became fairly acute in the middle and late nineties as broadband networks of all sorts began to court business and residential customers and to compete with one another.

The need for OSS is even more acute today because the range of service offerings has increased as has the ability of broadband networks to support differentiated services, and because of the great expansion of networking protocols now in common use and the need for a means of communicating across protocols and platforms.  Add to that the competitive pressures posed by multiple broadband service providers in the same local marketplace, and the choice of OSS products becomes just as important as the basic hardware occupying the central office and the network points of presence.

It must be said, however, that the pressure on wireless broadband operators to adopt comprehensive OSS software suites has been less intense in the recent past than has been the case for cable, DSL, and optical fiber network operators.  That's because broadband wireless has come to be identified with basic Internet access within rural or in less developed urban markets.  Notes Kevin Suitor, vice president of marketing and business development for Redline, one of the few hardware vendors also offering OSS, "the situation in the past, at least in the U.S., was that you had four or five thousand disrupters adhering to a WISP model, not a full service network model.  They were able to manage their networks with Xcel spreadsheets, but unfortunately that approach doesn't scale."

Plainly, all you can eat, basic access services with no performance guarantees are relatively easy to administer, especially if the operator has installed a one brand network, which is often the case in less sophisticated markets.  It's only when the network operator enters tough competitive environment requiring tiered services, voice, multi-media, and stringently enforced service contracts that the need becomes urgent.

Nevertheless, almost no one believes that the basic access service model can prevail indefinitely, and, since voice over IP already figures prominently in most WiMAX and proto-WiMAX networks in operation today, the need for real OSS exists now.  Consequently, some, though not all OSS software providers are already soliciting and selling to WiMAX operators.  And at such time as more advanced services such as video programming, video conferencing, and VPNs become the norm among broadband wireless operators, OSS will become an absolute necessity to administer a network, and any operator who hopes to stay in business will be obliged to understand OSS product offerings in considerable detail.

The change from circuit to IP protocols has also accelerated the development of OSS.  "I'd call it a revolution," says Mark Nicholson of Syndesis.  "There's a lot of legacy OSS out there developed for circuit networks, but the new, sophisticated services available only over IP networks have really been a stimulus for OSS development."

The Lack of Specialized OSS Products for WiMAX Operators

OSS is currently a fragmented market with no one vendor providing a comprehensive solution.  Instead a best of breed approach prevails, although some vendors provide more than one type of software product.

Aside from a few of the WiMAX equipment vendors themselves, no company is aiming a product specifically at WiMAX operators.  As Dan Baker, a principal analyst with Technology Research Institute, observes, "you see one or two small players at trade shows attempting to court broadband wireless operators, but generally the industry has ignored them."  The assumption instead is that since WiMAX networks share the same IP core as cable, fiber, and DSL networks, they can make do with any product aimed at automating services across those cores.

To a certain extent this is true.  There is nothing in the nature of WiMAX networks that would absolutely preclude the use of generalized OSS products, but in practical terms the fit of such products is often poor, because, among other reasons, they are aimed at service providers boasting thousands rather than dozens or hundreds of customers and are priced accordingly.  OSS vendors are accustomed to closing multi-million dollar sales with large network operators, and, to many, the hand holding required to service a small operator is not worth the relatively paltry revenues to be had from automating a small network.  Nevertheless, most of the companies contact for this series indicated an interest in meeting the needs of WiMAX operators, if not now, then in the future.

On this point John Armstrong, director of product marketing for Intelliden, a leading OSS software provider, remarks, "there's no question that our target market is the service provider commanding tens of thousands of devices, not a few hundred, but we are pursuing WiMAX accounts.  In the core of the network, the management problems and the utility of our solutions is the same regardless of physical layer."
Mark Nicholson, CTO of Syndesis, a major OSS vendor, concurs, and adds, "the attraction that WiMAX holds for OSS vendors has more to do with its potential than its current state of development.  WiMAX could conceivably break the duopoly of cable and twisted pair copper, and to us that's exciting."

Still, such vendors may find WiMAX networks considerably more challenging to serve than their wireline counterparts because it is difficult to utilize general purpose OSS software within WiMAX networks in order to provide different services and service levels simultaneously over limited bandwidth.  Observes Dan Baker, "the air interface requires special provisions that won't be present in OSS designed for other physical layers."

Nicholson is more specific, "the fading airlink sharply limits the available bandwidth for supporting value added services and creates a dynamic environment offering low determinacy when precisely the opposite is required.  It's our opinion that existing OSS won't be able to deal with WiMAX.  The systems will have to evolve to meet the requirements of these new networks."

Here it should be mentioned that the whole idea of differentiated services and quality of service began in the backbone of telecommunications networks back in the late nineteen eighties when the Bellcore organization serving the ex-Bell operating companies began the development of the quasi-packet system known as ATM (Asynchronous Transfer Mode).  ATM allowed the network operator to use a single physical channel containing virtual circuits consisting of predetermined arrangements of equal length packets known as cells.  The virtual circuits were supposed to exhibit the same determinacy as true physical "nailed up" channels while making more efficient use of available bandwidth.  At the time, variable length packet protocols were deemed incapable of supporting quality of service and delay sensitive traffic, and thus cells represented what was considered to be a judicious compromise between efficiency and determinacy. 

Whatever, the truth of the theories, the reality was that ATM networks generally ran over fiber and their performance depended as much upon the ability of fiber to provide users with stable, high bandwidth connections as it did upon the functioning of the ATM protocol.  While attempts were made to adapt ATM to wireless networks, the lack of resend mechanisms and the requirement of the protocol for maintaining a handshake across the devices in the network made such adaptations highly problematic.  In the days when QoS was synonymous with ATM, QoS was simply unavailable to broadband wireless operators.

Since the now distant days when ATM was seen as a universal transport protocol, many attempts have been made to develop equivalent QoS mechanisms for variable length packet networks, and some of these involve prior allocations of bandwidth that are in effect virtual circuits and are often so designated.  While more sophisticated than ATM, they still involve the same fundamental tradeoff-network determinacy versus efficiency.  Determinacy is no better than was the case for ATM, but efficiency is generally superior, so we may be said to have made progress. 

The problem is that such systems begin to break down when bandwidth is limited, and bandwidth is nearly always limited relative to demand for it in WiMAX networks.  It becomes hard to reserve bandwidth for a single use when there is an urgent momentary demand for it for another application.  And the situation is made worse by the fact that the airlink itself varies dynamically according to distance, the presence of interference or obstructions, and according to whether the user terminal is in motion, with maximum throughput rates fluctuating rather dramatically from moment to moment as a consequence.  All this tends to make traditional ancillary QoS protocols invoked through the OSS provisioning component much less effective than is the case with wireline networks where the effective bandwidth of the link is constant and predictable.

WiMAX has its own QoS mechanisms and will support all of those mechanisms already developed for IP, but most networks are still challenged in regard to bandwidth, and assumptions that hold for wireline networks and are generally reflected in OSS provisioning systems may not entirely hold in the wireless sphere.  We believe that this particular issue is central to the successful implementation of OSS in WiMAX networks, and will remain crucial for the foreseeable future.

OSS - Product and Market Segments

OSS software products reflect the complexity of the network and the range of its operations, but the basic product categories haven't changed in some time.

Network element management software embodies the core functionality that make network automation possible in the first place.  Devices such as routers, switches, gateways, application servers, and the radios themselves are normally configured through command lines conveyed through a terminal, and each has its own internal operating system.  Almost all current network devices can now be configured remotely and through machine to machine communications.  The element management software included in each device makes this possible.

The problem is that little standardization exists in the telecommunications industry for element management software, and devices performing the same functions but provided by different manufacturers will, in all likelihood, not utilize the same or similar software and thus will require different command lines in order to be made to execute the same functions.  Furthermore, a server containing a customer data base which has recorded a recent modification of a customer's service bundle has no inherent ability to command a router to provide bandwidth or to control latency in order to support the various services.

That takes something else, what is known as network management software which normally includes a mediation function.  The purpose of such software is to permit what is known as flow through provisioning, a process where the addition or deletion of a subscriber or a change in any subscriber's service bundle will automatically initiate a series of commands invoking appropriate adjustments in network devices from element management systems.

Explains Intelliden's Armstrong, "there are different ways of integrating diverse devices.  The way we do it is to model each device so that each command now occupies an abstraction layer which makes for a simplified interface.  All of the functions of each device are absorbed into the abstraction layer.  The other approach is to utilize canned scripts for the network elements where no real modeling occurs."

The next major grouping of OSS software is known as provisioning or service activation, and it is absolutely central.  Provisioning handles any changes in a subscriber's service offering and initiates the sequence of changes across the network as well as recording the subscriber's service status in a permanent data base.

Discovery is the term for crucial function performed by yet another class of OSS software.  Discovery refers to the act of discovering what network resources are available to enable a particular service offering for a particular subscriber, and thus discovery provides a necessary check on the provisioning function.  Essentially, discovery takes inventory of network devices and their capacities and then determines how much bandwidth can be provided and what service levels can be supported.

The last core component of OSS is known as assurance and whose purpose is to determine that the changes initiated in the network in respect to subscribers have actually been carried out.  Assurance may be seen as software based, fully automated quality control and provides the final determination that all of the components of the OSS software suite are fully integrated and working correctly.

Billing is sometimes considered a component of OSS, and sometimes not.  Billing software is generally the province of specialist firms.

OSS Integration

Stitching together the diverse elements of a comprehensive OSS suite can be an arduous undertaking.  The details of the integration process are covered in the second article in this series.  A number of common protocols and platforms are extant in the telecom industry to enable to different software components to exchange information, and, of course it is the function of the overarching network management software to bring order to the total ensemble, but because no protocols or platforms are universally utilized, it may be difficult to make the various software components operate harmoniously with one another. 

No single OSS software provider makes a complete turnkey system, though Oracle which recently acquired MetaSolv, comes closest, and accordingly the companies seek partnerships with one another and work in consort to make sure that their products will mesh.  Obviously, one needs to know the extent of an OSS provider's partnership before one purchases any software product.

Selection Process

Because OSS software selection does not represent a single choice, the network operator needs to have an overall strategy for the selection process.  That strategy will differ according to what if any infrastructure is already in place.

Ideally OSS should be included right in the final business plan for the network and decisions should be made before any hardware is purchased, in other words, OSS software should be seen as integral building blocks for the entire network.  OSS should not be grafted onto the network at a later stage, but should be there from the beginning.  Any attempts to operate the network with manual systems for configuration will result in recurrent labor costs that will quickly exceed the entire capital cost of the network and will constitute a profound impediment to adding services in the future.

All components of OSS are important, but network management software is primary since it integrates all other elements.  Provisioning and discovery are the key elements in handling interactions with customers while mediation is the glue that holds the OSS components together.  The choice of radios is arguably more important than the choice of OSS since the radio determines the capacity and availability of the network and its ability to support a range of services, but no radio should be purchased absent a consideration of how easily it lends itself to software control in the service of overall service objectives.

Adding OSS to an existing network, especially one containing legacy network elements, is a much more difficult undertaking, one that will require extensive consultation with OSS vendors.

Summing Up

OSS should never be considered an afterthought.  To a large extent OSS determines the overall performance of the network in meeting customer needs and has a major impact on operational costs and ultimate profitability.  Finally, it should be understood that OSS is still evolving, and that relatively little has been done to meet the specific needs of the WiMAX operator.  But even if OSS has yet to be optimized for wireless networks, it remains indispensable.

About the Author

Dan Sweeney, Ph.D., is an award winning business and technology journalist with over twenty years of experience and more than ten thousand publications to his credit including a recent book on WiMAX which has gone through two editions. Dan has also worked as an industry analyst, successfully tracking technology and market developments in the fields of energy, telecommunications, consumer electronics and transportation. He has particular expertise in systematic technology forecasting and he has written extensively on the subject. He has performed technology and market assessments for venture capitalists, angel investors, and venture backed startups, and has appeared as an industry expert at trade conventions and industry conferences.