The arguments for supporting open standards in the computing industry are well documented and include the need to appease user fears about being locked into proprietary systems. A less altruistic but juicier carrot spurring many vendors into co operation is the prospect of a vastly bigger market, stimulated by the introduction of open products. Network […]
The arguments for supporting open standards in the computing industry are well documented and include the need to appease user fears about being locked into proprietary systems. A less altruistic but juicier carrot spurring many vendors into co operation is the prospect of a vastly bigger market, stimulated by the introduction of open products. Network management is no exception to the trend with an estimated market value of hundreds of millions of dollars over the next five years. But the complexity of the subject, which embraces a myriad of functions from the basic detection, location and healing of faults to the collation of statistical and financial information being passed around a network, has to date relegated it to the backwaters of standardisation efforts. This now looks set to change. Entente A series of bi-lateral of talks mushroomed last month into a full-blown entente with the following agenda: to agree a set of network management options at each level of the seven layer open system interconnection stack, with the overall objective of knocking two years off the timetable for delivering open products that interwork onto the market (CI No 981). But while the eight strong OSI/Network Management Forum line-up includes the likes of AT&T, Northern Telecom, Amdahl and Unisys, the abstainers are even bigger. IBM and DEC have turned down the invitation to join, condescending only to chart the progress of the forum, and promising support at a later date if they approve the objectives which unfold. Both feel big enough to compete with the forum, IBM with its NetView product, which it is set on promoting as a de facto standard, while DEC is unwilling to throw away the considerable investment it has made in the field. Less obvious but perhaps more divisive, however, is the imbalance in representation between computer manufacturers and the telecommunications industry. A key factor propelling the forum is the different focus on network management adopted by the data processing and telecommunications industries and the need to merge these two perspectives, especially now that Integrated Services Digital Network is on the horizon. Cees Lanting, Hewlett-Packard’s European network standards co-ordinator explains that according to the data processing world view, network management governs those communications modules porting applications, simply because maintaining compatability between different end systems is paramount to the computer vendor. In the telecommunications camp, by contrast, attention homes in on the first three layers of the Open Systems Interconnection seven layer model, which are concerned with the nitty gritty task of shunting raw data about. These different views are further perpetuated by the standards-making bodies, the International Standards Organisation and the CCITT which champion the data processing and telecommunications cause respectively, in Lanting’s view. The distance between the two views must be bridged agrees Dave Taylor, research analyst with Dataquest. But so far the Bells and European PTTs, bar British Telecom, appear reluctant to support the forum. Roger Nucho, director of standards at Bell Atlantic, who is sceptical about the Forum’s objectives and charter believes it is better to devote resources to working these issues through existing standards committees to avoid duplication and confusion. He is also concerned about a clause which binds all members to implement emerging standards. Nynex probably speaks for its siblings in saying we support the implementation of global standards but cannot commit any support until the forum’s activities are better defined. While such caution is understandable – the telecommunications industry has traditionally had the edge in network management where the mastering of applications like billing is its life blood – network management is arguably the most critical issue currently facing the corporate networking arena. Industry today cannot survive for long if there is a fault on the network affirms Taylor. Crucially This sense of urgency has s
urfaced only recently but even now industry commentators agree that network management does not receive the attention it merits. Cost is one, not insignificant, deterrent. Software must manage every product that a supplier sells. The argument against developing open versions is even weightier – Crucially, every manufacturer has to make daily decisions on whether to expend huge sums enriching a proprietary architecture or whether to throw it out and start from scratch says Hewlett’s Lanting. If the latter path is taken, suppliers must bear the consequences of plunging their customers into a lean period where levels of net management are reduced. Because resources have been concentrated on specifying the more accessible features of Open Systems Interconnection, proprietary management systems are ahead of open versions, largely in terms of the user friendliness of products. Lanting claims that a rough specification has been defined right across the seven layer stack although some areas are patchy. Security is one notable gap, while configuration and performance management are still in embryonic stages. Forum members stress that the most important aspect of the new union is the pledge to implement open protocols and agree a time table for their products to interconnect. A target date for a demonstrator to prove inter-operability has already been set for 18 months’ time. Those suppliers with their own network management architectures, such as AT&T with its Unified Network Management, and Hewlett-Packard and Northern Telecom, jointly backing OpenView, have promised to migrate to the new standards as they are agreed. If the absence of some of the big shots renders the new union fragile, the intentions remain good.