 |
|
|
by Steve Broadhead Table of Contents INTRODUCTION:�
Still Making the case for Proactive Network Management INTRODUCTION: �still� making The Case for Proactive Network Management� Some things never change. Like many people�s attitudes to age, sex, drugs, rock & roll� and network management. For too many networking professionals it still means the software that is thrown in with the Ethernet switch that lets you configure it. That and the HP OpenView console running SNMP-based management � usually translated as that dusty UNIX workstation in the corner of the machine room with red icons glowing all over it, making endless .wav alarm sounds� Some people even claim they don�t need to manage their networks. That downtime is a regular feature on these networks is simply something they have always taken for granted. After all, it provides an opportunity to go for another coffee and chat about last night�s TV viewing. Doesn�t look so good on the weekly reports for the board of directors however. What�s that age-old argument about IT actually being counter-productive? The problem is that there is often a short-term, �cheap and dirty� solution to network problems: aka �throw more bandwidth at it�, usually more 100Mbps or even Gigabit Ethernet ports these days. The reason this methodology is a problem is that, in the longer term, it usually only creates greater problems � �moving the bottleneck� is a popular networking game with no ending � so it ends up actually being very expensive and very dirty indeed. It is also only applicable to simple LAN environments. Scene in the boardroom Technical support administrator requesting more budget for network management products from the IT director, who comments� �Network management, administrator � that�s a support function isn�t it? And support functions � well they�re non-profit making so they�re low priority. No, instead focus on making everything go quicker � buy a load of that gigabit stuff. Goes incredibly fast I hear.� �But it�s the field engineers we�re trying to support and gigabit�s not yet available over GSM sir�� Even where network management is being taken seriously on the LAN, often it is sadly lacking elsewhere within the company � often at the branch offices where arguably, because of the remote location and typically less technical level of staff, management is more than ever required to be in place. There are � as we�ll discuss shortly � valid technical reasons as to why this might be the case, but it need not be so. A more fundamental issue is that, despite regular reports from a variety of market research and analyst sources � check out the Gartner Group site and newsletter - showing just how costly network downtime is to each and every company and not just the �e� service providers, still very little is done to make the network more reliable in the future. Instead of looking to be proactive in managing the network, it�s a case of fire fighting when problems do arise. Any budget spent is typically done so on improving resilience in the hardware � many companies still buy two of everything and have a complete set of hugely expensive, redundant hardware waiting patiently to come online in the event of a problem. However, when the problem arises it is often unrelated to any of these, well-engineered, resilient areas of the network and there is still no methodology in place for resolving the problem quickly, let alone making sure it didn�t happen in the first place. A study from last year concluded that 92% of network managers say they are unable to guarantee service or manage their e-business network satisfactorily because of the issues of too rapid growth and the resulting complexity of their network infrastructure. Ninety-two per cent!!!!! So you think your network is managed? Does this scenario sound familiar? �Let�s run through those management features once more.� �SNMP?� �Check.� �RMON?�� �Check.� �OpenView?�� �Check.� �Sniffer?�� �Check.� �CiscoWorks?� �Check.� �Optivity?� �Check.� �Transcend?� �Check� �Sorted. So we�ve got network management. No one can say we haven�t. Look at all these tick boxes we�ve got checked. Sorted.�� �So why is the network down then?� �Well, er� has anyone been down in the machine room to look at the OpenView, CiscoWorks, Optivity or Transcend screens recently? And did anyone enable the pager alarm option? Mine doesn�t seem to be working. Yes I know the whole of the call centre is down currently. Yes, I know it�s costing us big time� but it�s not my fault that the network management isn�t managing the network. Does anyone here actually know how to use a Sniffer? No? Oh well, better bring in those exorbitant consultants again�� While network management is about monitoring the network to identify faults as quickly and accurately as possible, it is also about constantly analysing overall network behaviour and planning ahead to make sure that classic problems such as network segment traffic overloads simply aren�t allowed to occur. Proactive network management in other words, which does add value to a company and, in the case of major network failures (and there have been plenty over the past few years) makes the difference between still being in business and being another business statistic of the wrong kind. The Cost Justification Process: Proof of Concept Gone are the days when justification of an IT purchase extended no further than: �it�s the latest, greatest, fastest, newest, biggest, smallest (delete where applicable) device on the market � let�s get one, no two to be on the safe side.� Now � and rightly so � the business benefits of IT expenditure need to be understood and proven before any acquisition can take place. Cost justification and quick ROI (Return on Investment) are the order of the day. So before we even evaluate the quality of the Chevin TeVISTA product reviewed here, we need to see, even if it lives up to Chevin�s claims, whether it makes financial sense or not. ROI used to be measured in years but now it is more likely to be months if that proposed expenditure is going to be approved. So not only must an investment in IT provide bottom line business benefits, but it must do so quickly. From a technical standpoint, one of the key claims of Chevin is that TeVISTA enables significant savings on downtime to be achieved � the reasons for this become clear later in the report. Now, we all know just how expensive downtime is in this 7x24x365 world so many companies now inhabit. However, we need to be a little more scientific than to simply say that, if TeVISTA delivers what it promises then you save "shedloads" of money. The figures we�re using here come from a report recently published in the US by Newport. In the report they concluded that downtime is more than an inconvenience - it has a quantifiable (and qualitative) impact on productivity, revenue and employee morale. For companies transacting $1 million of business daily, downtime costs are about $9,500 per hour for B2C focused companies and nearly three times that per hour for B2B focused businesses. We have taken the far more conservative B2C related figures and applied them to downtime figures reported for various sizes of organisation yet to equip themselves with TeVISTA type functionality.� The assumption is 48 hours a year for small companies, 16 hours a year for mid-sized companies and eight hours downtime a year for systems within large enterprise companies.� In our opinion, the addition of TeVISTA is likely to reduce trouble-shooting time by 50%., There are a large number of other associated costs with a failure to manage your network properly. These include excessive use of costly, external consultants, unwise and unnecessary acquisitions of more networking hardware, training, inefficiency of staff unused to new systems� the list goes on and on. Downtime, however, is the real killer � literally meaning the death of a business in extreme cases � and should be avoided, literally, at all costs.� The example above, with Chevin�s TeVISTA, looks to be a very cost-effective form of insurance indeed, especially when you consider that the primary objective of the system is to avoid downtime in the first place. Network Management: The current technical State of the Art For many networking professionals, network management still revolves around the dual standards of SNMP and RMON. Hardly surprising when you consider that just about every networking device � and many PCs � are supplied with SNMP MIB II support as standard. Most � though by no means all � Ethernet switches also come with RMON support integrated, albeit at a minimal level in many cases. So it�s a case of �in your face� management standards. SNMP is ubiquitous. The problem is that these standards are technically flawed and inherently limited in their ability to manage any network, let alone the networks of the new millennium. Just because they are omnipresent doesn�t mean that they are actually God�s gift to networking. Nor is it that companies like Chevin, whose TeVISTA product we�re examining here, are able to work exclusively outside of these standards. Instead they are partially confined by them in the name of open standards and interoperability. So it is very much a case of providing standards-based products with as many improvements to those standards as is realistically possible to add. Notice how many times the word �standards� keeps cropping up� We�ve stated this before but, in the view of the NSS Group, there has been far too much �lip service� paid to SNMP and RMON support with too little effort made to provide management tools that � while supporting these standards � do move the science of management forward. That Chevin is attempting to do so, albeit within the constraints of industry standard support, is a positive step. No one wants to go back to the old days of being locked into a specific vendor but, at the same time, simply adhering to �lowest common denominator� technology in the name of standards is not good enough either. SNMP was borne out of the philosophy of freedom of choice but was never the best technical solution around. Even so it became the absolute, internationally accepted standard for managing network devices. And it was followed by RMON (meaning Remote MONitoring, not an abbreviation for that classic blues instrument), effectively a specialised SNMP MIB intended to provide base level statistics taken from nine statistics groups, designed to inform the network manager of the absolute status of any device. Being entwined with SNMP gave RMON a status advantage over any other mechanism in that it was immediately acknowledged as an international standard. And when measured in terms of support, RMON has undoubtedly been successful, as pretty well every network hardware vendor incorporates it within at least some their offerings. This, in turn, gave the user-base the industry standard solution it wanted. However, as is so often the case with standards that are agreed between vendors with conflicting proprietary interests, RMON � like SNMP - is a compromise solution that falls short in a number of very important ways. For example, it is close to impossible to find an Ethernet switch that supports all nine RMON groups simultaneously and continuously, because of the huge demands it makes of the hardware. Instead most switches support, at best, the first four RMON groups per port and only one port at a time can be monitored using all nine groups. Absolute network visibility? Don�t make me laugh� RMON shortfalls The amount of information that it provides is not complete enough for network managers and administrators who are faced with resolving complex problems at a distance, a problem the Internet explosion and the increased mobility of workforces has exacerbated. Despite being a defined international standard, a very large percentage of network hardware vendors do little more than pay lip service to RMON support, in the sense that they implement only a fraction of the standard. Typically they will support the first four of the nine RMON groups � sufficient to be �RMON compliant� without having to invest in the kind of hardware levels required to effect a complete RMON solution. The mechanisms employed for data retrieval to the central SNMP management console approach - typified by something like HP�s OpenView - are slow and very bandwidth inefficient. A major specific problem with RMON is that the data �conversations� are not stored as bi-directional structures. Instead they are stored as �A to B� then, elsewhere in the tables, from �B to A�. This means that - even if the two elements required to build a true picture of a conversation are adjacent in the table - there will be an inherent latency between the �A to B� and the �B to A� retrieval. This results in an imbalance between the values, so the stats are not accurate. The inefficiencies of RMON are well documented. For example, on a 200 node LAN, you would expect typically to see at least 1000 conversations at any time. Two requests per conversation, each requiring a response, means that RMON must generate 4000 packets to see all the conversations. In other words, it is creating a huge amount of traffic merely in order to manage the network. Ironically, it might just as easily grind the network to a halt in its attempts to do so. While RMON - and SNMP especially - are here to stay in the mid-term at least and therefore need to be supported, network management needs to offer far more than just the basic levels of support for these standards, in order to be of any real use. Moreover, there are definitive products � market leaders� - in all areas of network management that are now so engrained within their user base that they also need to be supported as if they were industry standards. What is required then, is a new approach to network management that is designed with modern, primarily (but not exclusively) IP-based, networks in mind, whether they are office-centric, remote, eCommerce or eBusiness oriented. If, at the same time, the existing network management standards and popular components need to be supported then clearly this presents a challenge to the vendors. Chevin, for one, believes it has an ideal solution that doesn�t force any kind of compromise, in the form of TeVISTA. So let�s now put it to the test. The Chevin technology approach outlined Chevin�s approach to network management appears to be quite simple: make sure you can see everything that�s happening on the network, wherever that network might physically reside. In this sense, the focus on pulling management data in from all corners of the network, especially remote locations with low-speed connections to a head office network, is what differentiates its approach with TeVISTA compared with classic SNMP implementations. The distributed network management challenge In a world where there is just one big computer - say an IBM mainframe - regardless of the number of terminals attached to it, the concept of managing such a system is within the grasp and capabilities of most individuals. This wholly centralised approach to computing makes the management of it relatively simple. However, it also makes for a very rigid framework, one at odds with the mobility now available to companies and their staff. Hence the continued emphasis on networking. But even with the domination now of IP, so reducing the multi-protocol nightmares of a few years ago to a great extent, networks are complex beasts to manage compared with their mainframe counterparts. It's not as if the network you are looking to manage is even restricted to one physical location. The chances are that it is spread across 10's or even 100's of locations, especially if you are in the retail or commerce business, �e� or otherwise, with widely varying connection speeds. The combination of growth in remote networking and the ultimate evolution of �client-server� computing, with data distributed across several geographically independent but technically dependent data centres, is changing the management picture once again. If companies are to realise the benefits of this kind of distributed networking they must have a strategy in place that is broad enough and flexible enough to make it manageable. Tasks carried out on distributed systems are often inter-dependent on each other so remote network management is now a critical requirement. This kind of sudden need for a management tool, which hitherto has been deemed unnecessary, is typical of the constant change network management is currently going through and will continue to go through in the foreseeable future. The old motto, �be prepared� springs to mind� At the heart of the technology designed to enable TeVISTA to gain total visibility of a widely distributed network are two elements, the Visibility Agents (VAs) and HSRmon (High-Speed RMON). The VA is a software agent that, in a sense, can be likened to an RMON hardware probe. It sits on a server or PC client and reports back to the TeVISTA central console, in the same way that an RMON probe reports back to an SNMP console. The differences are that, firstly, the VA is very inexpensive and easy to deploy across the entire network � one per segment is all that�s required to get total visibility - and secondly, wherever that part of the network is located and whatever the speed of the network connection, the VA can accurately report back to TeVISTA. The reason why the VA is able to do so is due to the second element � HSRmon. We�ve already emphasised how the inefficiencies of RMON and SNMP�s bulk data transfer methodologies create excessive traffic. This, in turn, makes them somewhat less than ideal � useless, basically � for managing remote networks across low-bandwidth connections. For this reason Chevin created HSRmon, which is designed to absolutely minimise the amount of traffic created in order to manage a network � remote or otherwise � in real time. Returning to the earlier example of RMON generating 4000 packets to view 1000 conversations between 200 network nodes, reveals that the equivalent conversation matrix update built by Chevin�s HSRmon uses less than five packets. Which is at least 3995 less that RMON took to achieve the same means to an end� In a performance context, this means that on a 10Mbps Ethernet segment, assuming a realistic 1ms RMON probe/agent turnaround time, it will take four seconds to complete the request. At the same time it will have generated 16% extra loading on the very network that you are attempting to protect from excess traffic. Take this into a WAN or Internet environment and the effects are considerably worse. Performing a simple bandwidth equation, and forgetting that the WAN or Internet link will also be required for other services, shows that it would take in excess of 10 minutes to transmit this data across a 64Kbps ISDN line, for example. And all for one, probably insignificant, item of data. Even in a LAN environment primarily based on switched 10/100Mbps Ethernet connections to the desktop, the efficiency of HSRmon is valuable. With feedback from vendors showing that some clients have reported up to 70% of their LAN utilisation being generated by network management traffic, maybe HSRmon�s role is actually invaluable. However, it is the other element � the VA � that offers the really invaluable service on a LAN. Where RMON fails spectacularly is in offering complete visibility of a switched network, designed as it was for a single-segment, shared LAN environment. Within TeVISTA, a combination of the VAs and HSRmon gets around this problem by enabling multiple VAs to be aggregated and treated as a single entity, so that a complete switch � each and every port � can be managed, real-time in the same way that RMON monitors a single switch port. HSRmon then takes care of delivering what would be an incredible amount of data under RMON as efficiently as possible. As a solution to the management requirements of contemporary (and future) switched, distributed network environments goes then, TeVISTA appears to have been engineered better than most. But how easy is it to actually use? Read on� TeVISTA is a PC software package consisting of the primary management application, the Visibility Agents and a number of extra utilities, including a network traffic generator, direct drivers for PCs using 3Com�s 3C5x9 Ethernet NIC family and user guides in html and Microsoft Word formats. The TeVISTA application runs on any of the Windows variants from �95 and NT4 onwards. The minimum recommended PC specification is a Pentium 450MHz machine with 128MB RAM. The VAs support the same range of PC operating systems and use around 6MB RAM. Any NDIS driver-based NIC is supported, in addition to the direct drivers for 3Com NICs. In terms of network support, TeVISTA is aimed at any network consisting of any or all of 10/100/1000Mbps Ethernet, 4/16Mbps Token Ring and FDDI � basically anything bar ATM.� Capacity wise, Chevin claims that the product will happily support networks up to 10,000 nodes in size currently, but makes the point that there is no theoretical limit to the system�s capacity. Though essentially a single product, TeVISTA is very modular in structure, based around a hierarchical model with a top level Enterprise Manager providing an overall view of the total network. Below this is the Network Asset Manager (NAM) which identifies individual network components and from which the majority of TeVISTA processes are fired up. These processes include the Total Visibility Management (TVM) tool, packet capture and protocol decode tools and report generators. For the latter, TeVISTA requires Microsoft Office �97 or later to be installed on the PC hosting the Chevin application as it auto-generates reports in Excel and Word, as well as html, formats. Just how you view TeVISTA depends on which of the various elements within it are of the greatest value to your particular set of requirements. In one sense it can be viewed as an SNMP-like network mapping and fault-detection alarm system, yet in another sense it�s a network trouble-shooting tool-kit. Equally it could be described as a network status and service level reporting application, a network health check report generator or a trend analysis and network planning tool. In truth it is all of these things and more. This, ironically, could be a problem for Chevin if, in a pre-sales situation, it fails to map the requirements of what could be a very varied potential customer base onto the specific elements of TeVISTA that particular customers are interested in. We say ironic because the strength in depth of the product�s feature set is really a huge bonus. It�s quite likely that very few customers will actually use all of the features within TeVISTA but that is equally true of many networking products historically. What is important to understand is the differentiating factor of the product � in this case the HSRmon transport protocol and the visibility agent combination. No other network management product we know of has anything similar. TeVISTA is supplied on a single CD with � a rarity nowadays � hard copy manuals accompanying the software. The CD autoloads and presents you with a choice of installing the software or browsing the CD contents. Selecting the latter, presents you with the option of installing the LoadLan traffic generator, Winsock 2.0 support and 3Com direct drivers and online copies of the user manual in Word or html format. Choosing the former brings up a wizard-based install routine for TeVISTA which quickly installs the software. Before you run it for the first time however, you need to install the supplied security dongle by attaching it to the parallel port of the PC hosting TeVISTA. The dongle is an item that Chevin has always insisted on using with its software in the past and this has been retained despite ongoing debates and criticism from some quarters � mainly NSS! � as to its use. In truth, however, the dongle does keep customers honest, has never proved to be unreliable at any time and does mean that the TeVISTA data on your network is secure. The second aspect to the installation is the loading of Visibility Agents. The PC hosting TeVISTA automatically loads a VA as standard but � depending on the licensing agreement � VAs can be loaded onto any number of PCs on the network. The basic license supports 25 VAs which are coded to match the dongle. On loading up TeVISTA for the first time a VA installer can then be created to be distributed either on floppy disk (or CD/RW) or via a networked drive so it can be installed remotely. Bear in mind that you need one VA per network segment in order to achieve complete visibility. Once installed, the VA loads automatically (it runs as a service under NT4). NB: The LoadLan traffic generator utility does not require the use of the dongle and can therefore be installed on any PC on the network. However, it cannot be loaded on any machine that is hosting a live Visibility Agent. We installed TeVISTA on an HP Vectra Pentium II 400MHz PC with 160MB RAM, running Windows 2000 Professional, thereby complying with the basic specification guidelines without the need to use a more recent PIII or IV machine. The key point here is that, unlike many management products, an expensive high-specification (often UNIX) system is not actually required in order to run the software. The NIC used was a 3Com Etherlink XL 10/100. Having loaded up TeVISTA we then created the VA distribution software and installed a number of visibility agents across the network. Obviously the amount of time it takes to install these depends on the size, topology and geography of the network, but in our test lab environment total installation time was scarcely 10 minutes, including loading the traffic generation utility onto a couple of PCs When you load up TeVISTA for the first time, the initial reaction is, ironically: �well, there don�t seem to be too many options here.�� We say ironic because how wrong you can be? Basically you are faced with a largely blank screen and a small number of menu options and an icon bar at the top of the screen.
The first of these options: adding a new item � an icon � to the screen, is the starting point. This is the absolute top level of your network so, for example, you may want to add a map showing the whole region the network covers � the world even � and specific icons for each branch of the network. TeVISTA supports OLE so it is possible to include background pictures and layouts, geographical maps for example, then open one application from inside another. For example, you may have a diagram showing the layout of your offices with the network map superimposed upon it. The other icons on the bar are primarily for zooming in and out of this top-level map of the network you create. On creating the new item you name it � in our case NSS � then double-click on the icon to get to where the action really starts, at the NAM (Network Asset Manager) level. At this point, two windows are open � the original Enterprise level and the NAM, so switching between the two is easy and they are available at all times. This is a general feature of TeVISTA. As you drill-down into the network and/or run the various sub-applications available, these typically run in a new window so each and every level of the network is always available without having to restart a particular element of TeVISTA.
At the NAM level you are presented � initially � with a similar, largely blank screen but a couple of important options on the menu/icon bar, as well as an additional set of tab options. The first menu option you will use is to add a new item � in this case where you define the network, or a subnet, and let TeVISTA discover all nodes on the network, whether they are switches, routers, servers, PCs, peripherals...You are given the option to specify a single IP address or a range of addresses, or even a named item. There are also a number of discovery options available for TeVISTA, as illustrated below.
If checked, TeVISTA will test the network node to see if that particular service is available, such as Telnet access for example. Within NAM, these options can be set as default scan and verify options, then edited for specific discoveries, as required. One point to bear in mind with large networks especially is that the more service tests you request, the longer the discovery will invariably take. We requested the discovery to check for all available services and � on what was primarily a 10/100Mbps Ethernet test network � found that typically TeVISTA took around 5 seconds to complete a discovery on a local 100Mbps switched node and up to 25 seconds to carry out the same process for a node on a routed segment via a 10Mbps Ethernet connection. Obviously these times can be reduced if less test options are specified. As TeVISTA discovers a node, it runs through the various service tests you specified, then creates a node on the screen. It attempts to identify exactly what the device is � for example, a 3Com Ethernet switch, a server, PC etc � and create a relevant icon from the supplied range of bitmap icons. These can be added to but include common vendors such as 3Com and Nortel as standard. Where an SNMP or VA is discovered, this is also specified by the type of icon created. Any icon can be changed manually afterwards if so required. Hovering the mouse above the icon displays all the information that TeVISTA has discovered about the node. This include IP address details, what services are available � such as Telnet access � and RMON and SNMP details, where relevant. Once an initial discovery has been run you can then perform as many extra discoveries as you want � for a different or new subnet, for example � from the second icon on the menu bar. As discoveries are run, TeVISTA stores the ranges you specified in a list from which you can also choose at any point in the future. It also serves as a reminder of which discoveries you have carried out to date. You can also �verify� the network � effectively re-running the discovery over a specific IP address range to check for any changes or additions. If found, these are flagged with an extra icon so you can immediately see that changes have occurred. The flexibility of the discovery options is very well thought out and would certainly prove invaluable over a long period of time.
With what may be thousands of nodes discovered, it is important to be able to arrange the network on-screen as you want to see it. With TeVISTA you can move the icons wherever you want and create multi-screen maps. You can also zoom in and out of specific areas of the network to view the onscreen details more easily. Once the network is laid out on-screen as you want it to appear, the icons can be �locked� in place to ensure that the network map is not accidentally changed. Just one problemo�. We did encounter one problem when checking a group of HP Vectra PC clients for SNMP agent support (which they did have running as a service under NT4) where the discovery mechanism would fail to get beyond a specific PC. After investigation we found the problem to lay with an old, memory resident version of McAfee�s anti-virus agent that auto-loaded onto the PC and prevented the SNMP agent from �talking� to TeVISTA, that came packaged with the PC (a �97 model Pentium II). Interestingly, a more recent version of McAfee�s anti-virus product on a Windows 2000 platform gave no problems. With the anti-virus agent removed from memory (by stopping the NT service) in the NT machines, the discovery completed successfully. We were then able to restart the anti-virus service. No other problems were encountered at any time during the testing. Configuring The Network Asset Manager (NAM) With the network discovered and mapped, you can now set about configuring the NAM to manage the network exactly as you want to. In addition to the menu and icon bar, there are three tabs labelled �Graphical Map�, �Pinger Controls and VA Controls�. The graphical map we�ve already discussed briefly but the other tab options are also important. Every IP node on the network can be set up so it can be pinged as required. Here you can set the ping frequency, in seconds between pings, what time out (in ms) to apply before failing a ping and whether to log ping events and create and view a .csv file of response times (in Excel for example). Alarms can be enabled and configured based on carrying out a particular action � such as a warning sound � if the ping fails. For each IP node you can view the responses real-time in graphical form.
In addition to setting up pinging controls, a third tab lets you set up the VA controls. This screen shows details of all the devices running VAs, with IP address and device name shown. Here you can set up statistics collection details for each VA. These stats then form the basis of a network health check report you can produce (more on this later). The stats are also saved for analysis off-line, so you can start to monitor network trends as part of future planning procedures. There are a number of stats options available, including specifying the sample period and number of samples per file, whether RMON conversations should be collected and defining how you want the stats to be collected overall. The stats collection can then be triggered from here and alarms appear in a box to the right of the Visibility Agent details, along with the status and overall stats details such as overall packets collected and overall network usage. It means that the Visibility Agent controls can equally be used for real-time or historical analysis and � most importantly � are really easy to use.
We�ve already mentioned that multiple VAs can be aggregated and treated as a single entity, so that a complete switch � each and every port � can be managed, real-time.
This really does give a total view of the network in a way that a regular SNMP/RMON combination simply cannot. Since they are only focusing on one switch port at a time, there is no way of knowing what is happening at all the other switch ports. However, with the visibility agents, you just assign one to each switch segment, then create an aggregated VA by combining these on-screen and then monitor activity on that aggregated VA. We know of no other product, currently, that is able to provide such a comprehensive view of the network, so easily. The network map can also be changed to show which PCs hosting VAs are aggregated, so that you can see exactly what the current TeVISTA/VA configuration is at any time.
Online help Throughout TeVISTA, online help is always available, as illustrated above, alongside the Visibility Agent aggregation illustration. One downside to the help screens is that they are not context sensitive, so finding out information always requires scanning through the help index. This can be both slow and laborious. On the plus side, however, the help is effectively like a tutorial, guiding you step by step through most TeVISTA operations, so little in the way of training is actually required in order to use the product. What Chevin could, perhaps, add to the documentation � online or otherwise � is a scenarios guide, highlighting a number of probable network problem scenarios and describing, step by step, how you could use TeVISTA to resolve those problems. Managing The Network With TeVISTA From the main network map screens within the NAM you can carry out a huge number of different operations. Just what you are able to do varies from one node to the next, depending on what services they have available. For example, if the device is a switch with RMON, SNMP, Telnet and web-based management support, TeVISTA lets you directly access any and all of these services directly from NAM, via a right-click of the mouse button and a selection from the drop-down menu that then appears.
The big advantage of this approach to network management is that it means you can directly manage a very wide range of different vendor devices from the one interface (see section: How TeVISTA can reduce the number of network management components). It also means that, in some instances and especially for SME customers, it may be the only network management product you need, other than helpdesk and similar peripheral utilities. Equally, for those with existing network management infrastructures headed up by something like HP�s OpenView, TeVISTA can effectively plug in below and give a far more accurate vision of the network while also being able to help resolve network problems, as well as helping reduce them in the first place. Taking each NAM function in order, we�ll now look at these in more detail. Any devices supporting SNMP can be interrogated using the SNMP Reporter. This can be used to open all MIB2 enabled interfaces on the device and view the contents, as you might with a pure SNMP-based management console.� Adherence to the SNMP standard at this level provides users with the ability to verify information presented by TeVISTA with that held by SNMP Agent software at a node level � useful insurance. � Any device that supports access via a web browser � such as a web-based configuration manager on an Ethernet switch � can be managed from here. TeVISTA automatically loads up the browser and points it at the relevant IP address/URL. As we�ve already explained � but it is an important point - since many network devices now are primarily managed via a browser, this means that TeVISTA can provide access to almost every vendor�s network hardware direct from the NAM, without loading up any other software. It supports multiple, simultaneous HTML sessions, so you can manage multiple devices at the same time. Arguably even more networking devices support Telnet access than web browser based access for management. TeVISTA is able to open a Telnet session directly with any device supporting Telnet access. As with HTML sessions, it can support multiple simultaneous sessions. Similar to the above, if your network device supports SMTP you can open up an SMTP session with it directly from the NAM. This is the key module, enabling you to drill down to particular devices to troubleshoot, capture packets, and decode them to view the actual contents. In addition all the statistical data can be archived, viewed offline, and converted into management reports describing the current health of your network, which we�ll describe later in the report. Any node with RMON support or a visibility agent loaded has a number of extra management features available to it, in addition to being able to directly access RMON statistics from within the NAM. Nodes supporting RMON gain access to TVM (Total Visibility Management) � either directly or via a VA � Trouble Shooter, statistics collection and offline statistics analysis modes. If you choose to access TVM via a visibility agent you gain access to the full range of statistics, via HSRmon. Nodes with VAs loaded further have direct access to Discovery and Packet Capture modes, which we�ll also describe separately. Having selected which VA you wish to use, on loading up the TVM module, you are presented with a new screen showing multiple different breakdowns of network activity, both graphically and in tabular format. There are six tabs listed Node Statistics, Conversation Statistics, IP Conversation Statistics, IP SubNet Statistics, Protocol Statistics and Graphical Map. Node Statistics presents a table of all discovered network nodes by MAC and IP address, and includes statistics on errors, a breakdown by packet and octet sent and received and the number of conversations the node has been involved in. Two graphs at the top of the screen show LAN usage and average packet size in addition to general information showing details of the time the software started, the current time, total packets and octet counts, node and conversation counts, broadcast and multicasts, and protocols detected on the segment. Chevin is keen to stress the �real-time� attribute of this data, claiming that the latency associated with SNMP makes it impossible to ever get a complete, up-to-date picture of nodal activity. The key upside for the user is that total network activity can always be viewed, for example when replicating user problems or issues.
From the Node Statistics screen, when highlighting a node in the table you also gain access, via the right-mouse button, to a sub-menu giving you the choice of tracing traffic to the node, from the node, or to and from the node. You can also list or view conversations, edit the node properties or simply check the node properties. The key point here is speed of operation � when you need fast access to any network data during a crisis, it is always readily at hand. From this sub-menu you can also open the Node Scope utility. This utility focuses on the LAN interface of the selected node allowing you to check - at a glance - the performance attributes of that device in relation to the LAN and, more importantly a WAN. You specify the uplink throughput and Nodescope provides you with a graphical view the expected utilisation of that link. It also has the value of being meaningful to non-technical staff. As well as showing overall usage it lists the top 10 users of that service, both currently and since the software was started up. It is also possible to run a packet capture from Node Scope itself, which automatically builds the appropriate filter. Conversation Statistics shows any or all the conversations between any two or all nodes, identifying each end node and its last conversation time, plus octet, packet and error details for each node. It also shows overall details of the time the software started, the current time, packets and octet counts, node and conversation counts, broadcast and multicasts, and protocols detected on the segment. In addition a graph shows segment usage and a bar graph shows packet sizes. From this screen you can also trace an individual conversation, which automatically invokes the packet capture utility. Getting back to basics Understanding the relationships between data flows travelling across a network is akin to understanding traffic flows on a road network.� If you do not understand why endpoint traffic flows exist, then it can be an uphill struggle to diagnose perceived network problems and implement an efficient network redesign. Witness the effect of bad road junction design and you�ll understand the potential problems that lie in wait. For most people pictures still speak louder than words, so fortunately TeVISTA presents the key information in obvious graphical terms, as well as in more detail beneath the surface. It means that both strategic and technical oriented staff can make use of TeVISTA by extracting that level of detail that fits their level of understanding. IP Conversation Statistics provides similar data to the above, but is restricted to IP traffic only, offering an opportunity to monitor activity across inter-networks, between sites or even across the Internet. It�s worth re-iterating here that one of the great benefits of the TeVISTA architecture is that it does allow you to make the most of low-bandwidth connections and manage across them without any compromise. IP SubNet Statistics lets you choose a subnet then displays the number of nodes active in that subnet and similar traffic details to the other screens already listed. The effectiveness of your network design can therefore be analysed with great clarity, as any fundamental limitations are glaringly obvious when comparing subnets. Protocol Statistics provides a graphical and tabular breakdown of network usage by protocol type. This is a useful facility, in ensuring that protocol activity doesn�t inadvertently affect network performance as a consequence of installing print servers or new Windows clients (NetBIOS), for example.
Graphical Map shows the network mapped with all conversations between nodes. Conversations are colour coded to show one or two-way conversations and whether errors are occurring on that inter-connection. This colour coding is used throughout the TeVISTA screens to denote potential problems or otherwise. By right-clicking on any node you can trace a one or two-way conversation, or open up Node Scope. Overall, TVM gives you access to all the features you need to discover and resolve problems, some of which are also accessible directly from the Trouble Shooter and Packet Capture submenu options from within the NAM (when right-clicking on a relevant node) so we�ll describe these more fully below. Another benefit of TVM is that its architecture means that you can really create a network map interface within TeVISTA that accurately reflects a visual representation of your their network. So when dealing with potential network problems, the logistics of TeVISTA make it that much easier to navigate your way around the network and resolve the problems quickly. The Trouble Shooter module is effectively a contemporary � and thoroughly comprehensible - version of ye olde network monitor. You can run it directly against an RMON compliant device, or via a VA for more efficient and broader statistics. It can also be used with stored statistics files for later analysis. Trouble Shooter encompasses a number of screens, some of which are in a similar format to that of Windows/NT Explorer, with a series of tabs to navigate through these different screens, as defined next.
The Explorer screen lets you view the network by individual node. Depending on the device, information may include the MAC address, the IP address, the device name, the group and type, the number of conversations and the first and last times it sent a packet. If the IP address has arrived via a router the MAC address of the router is shown. Right-clicking on a node address presents a submenu giving the option to view or trace conversations. A similar feature is also available on the conversation explorer and scope screens. The Lan Hints screen is used to inform about possible problems on the network. If a device has remained inactive for a long time it will be reported, or if the network itself appears to be unusually quiet a message appears suggesting a possible problem, such as check the cable for a possible break. Once a device sends or receives a packet it is reported as active again. Conv Explorer shows conversations between a selected node and the rest of the network at MAC or IP level. You can choose any node on the network to be the focal point, or let TeVISTA find the busiest node, or the node with the most conversations. As with other conversation screens in TeVISTA, colour coding is used to denote the conversation status and load status. Green indicates a two-way conversation with no errors, red denotes errors, blue highlights a one-way conversation (for example, a broadcast or multicast) and a dotted line show where a conversation has timed out. A particularly valuable feature of TeVISTA is that it allows you to isolate any node that requires immediate focus and check that users are actually �talking� to the nodes that they should be.� It can even be used to dynamically check on web activity should security or illicit surfing be suspected. Conv Scope uses a global view of the network to show all ongoing conversations at MAC or IP level. Since this can get confusing on a busy network, a slider control on the right of the screen controls the number of conversations being shown, from most to least busy, as the number shown increases. From a general business standpoint, the real value of this screen is in clearly depicting the complex nature of communications across a network and the need for network operators to keep tabs on cross-network activity in the interests of optimum performance. The Pie Charts screen presents a breakdown of network usage by busiest node, group or types in � you�ve guessed it � pie chart format. This is a snapshot of the network at the time the screen is brought up. A refresh button means you can always get an up to date snapshot and, in pie chart format, the overall picture can be seen and understood very clearly and easily. But altogether more interesting for us � if less pretty � is the LAN Wizard screen. This is designed as a fast-track method of identifying network problems. It consists of several sub-screens. Unresponsive Nodes lists, by name or address, every device not currently responding to other devices on the network, the number of packets received, the name or address of the sending device and the protocol. Nodes By Type enables you to list network nodes within a specific category, such as router, server or IP node, as defined in the NAM, so you can focus in on specific device types. It shows packets/octets sent, LAN usage and throughput in packets per second. Nodes By Group similarly lists nodes by group type as defined in the NAM. The Duplicate IP Address, Timed Out Nodes and Broadcasting Notes screens list all nodes fitting that particular description. The former, especially, can save a lot of time on a large network. On all these screens, highlighting then right-clicking on a specific node in the lists gives you the option of viewing or tracing a conversation to, from, or to and from that node. The important point to make here is that all too often you see network operators spend � let�s face it, waste - valuable time trouble-shooting network problems as a consequence of not having this kind of fundamental, and hugely important, information at their finger tips. The History screen provides a graphical record of activity on the network, starting from when TeVISTA was loaded and a snapshot of usage at a specified time. You can zoom in to peaks and troughs and change the time slice that you want to view. You can also refresh the snapshot at any time. The Data Lists screen displays details of all the nodes on the network in tabular format, with a number of subscreens showing details of conversations, IP conversations, subnets and protocols. By highlighting and right-clicking on a node details on any of these screens you can choose to view or trace a conversation. Exporting data At any time when using the Data Lists and the LAN Wizard, you have the ability to export data directly into the MS Notepad utility for manipulation in external reporting or spreadsheet applications. This is useful if you want to put together an instant, ad hoc report, for example. IP Device Browser presents a MIB tree for an IP address you specify from a pull-down list of qualified nodes. From here you can use GET commands to retrieve SNMP data. Any node with a VA active is capable of collecting and archiving detailed statistical information about the traffic on that network segment. These statistics files form the basis of the network health check reports�produced by the report generator, which we�ll explain in more detail later. In addition it is possible to view these files off-line later, so you can see how your network was performing at the time the statistics were collected. The collection can be configured by length of time and number of samples per file, as well as scheduled to run at a specific time, or continuously. You can also specify � device permitting � whether RMON stats should also be collected. From within NAM, on any VA loaded node, you can run a one-off discovery.
This opens up a new screen and lists each node as it finds it. Nodes are recorded by protocol detail, address and name. The latter, where appropriate, can be listed on the map as the IP, NetBIOS or Novell IPX name and, equally, salient data represented for legacy equipment such as mini and mainframe computers from IBM and Digital. This kind of point is significant as Chevin has designed TeVISTA to be a unifying management platform for IP and legacy systems alike, in contrast to most recent management software which ignores the world beyond IP. Right-clicking on a node details gives access to properties such as the number of conversations the node has been involved in, and packets/octets sent and received. As well as being available from the NAM level � depending on the validity of the network node being managed � packet captures can be triggered from many and various screens within TeVISTA, notably when in TVM mode. Before running the packet capture you can set up a filter to capture specific packet types, such as Internet or broadcast packets. A default filter is provided but you can also create your own. Instructions are provided in the documentation Chevin supplies to help you with this. You can also configure the capture in terms of the capture length - that is the amount of each packet to collect - and the number of packets that you want to capture. Filters can also be used to trigger the beginning and end of capturing. You may specify that the triggers activate immediately a packet of the required type is encountered, or you may wait until a certain number of that packet type has arrived before starting the capture. End Triggers may be set in exactly the same way as start Triggers. Otherwise it is very much a point and shoot operation, as indeed TeVISTA describes it - a case of hitting the Start button and away it goes. If a trigger has not been set to stop the capture � or even if it has � clicking on the Stop button ends the packet capture. At this point, if the default setting of �Start Decoder� when the capture stops is used, the protocol decoder automatically kicks in. This consists of a screen split into two windowed areas.
The left-hand window lists the packets in captured order with source and destination addresses and protocol type. Clicking on any packet brings up the packet details in layer by layer format, in hexadecimal and layers 2 thru� 7. At the top of the screen are controls for both navigating around the captured packets, �tape recorder� style, and for displaying exactly which data layers you want to display onscreen at any one time. This latter feature is really neat as it prevents you from being bombarded with unwanted data onscreen. For example, if you�re searching for a particular URL you can focus just on layer 6 information, or layer 4 if, for example, you�re searching for a particular port number. The main problem with the decoder is that there is no specific search or find function which would be very useful if you were searching for a particular text string, for example. You can restrict the packets displayed by certain criteria such as source and/or destination address, but it is by no means as useful as a true search engine would be. Usefully, packet captures can be saved and restored for analysis at any time from within the NAM, adding to the extensive trend analysis possibilities within the product. Evaluating TeVISTA yourselves � the �virtual showroom� If you don�t want to take the words of NSS for granted, or Chevin�s own publicity then there is another option that doesn�t require you to go through the process of a full on-site evaluation. Instead, Chevin offers what it calls a Virtual Showroom. By arranging an appointment with a pre-sales technical at Chevin, you can look at the product features, online across the Internet to get a fast and easy appraisal of TeVISTA with a minimum of effort.
While it may not be as conclusive an evaluation as living with the product on your own network for a few days it is certainly a million miles on from an online PowerPoint presentation! Trouble-shooting the NSS network Within the NSS labs we specifically set up a number of �problem� scenarios in order to test the trouble-shooting capabilities of the TeVISTA product on the NSS network. hile �common� problems such as duplicate IP addresses are easily captured by TeVISTA without the need to dig deep into the layers of the Trouble Shooter, we introduced more complex problems such as dealing with a denial of service attack to see how it coped. Using some evil public domain tools we attacked a web server from within the network, ensuring it was completely 100% utilised � server and LAN connection - and then put TeVISTA on the job. We were informed of LAN usage peaking and able to immediately identify the problem node � our Web server. We were then able to isolate the conversation and see where the source of the traffic was, without the need for words� At this point we were able to use the NAM to directly manage the Ethernet switch via its web-based interface and disable the port from which the denial of service attack was emanating. The whole process took literally a couple of minutes, from identifying the problem to resolving it. Obviously, the attack was somewhat contrived, but the basic problem of network saturation is one that is both very common and very varied in its sources, so the test was a valid benchmark of TeVISTA�s capabilities, we believe.
We�ve already mentioned that, wherever there is a VA loaded, TeVISTA can collect statistics throughout its operation, storing them as date-stamped files to be brought into TeVISTA for analysis whenever you want. This leaves you with a wealth of historical data to report upon for trend analysis, performance comparisons and similar tasks over as short or long a period of time as you wish to maintain the data. However, reporting is available at any time during operation, so simply by setting up a short statistics collection you can effectively get a �snapshot� data report almost �on the fly�. From within the NAM you can right-click on any VA-loaded node and choose a saved stats file � it may be one that has just been created, of course � to load into TeVISTA. At this point you are given the choice of running Report Generator or the off-line version of Trouble Shooter. Choosing the former presents you with more options � running an HTML or Spreadsheet based report (see Network Health Checks section), or viewing the Stats Browser.
The latter provides you with a simply formatted but comprehensive onscreen breakdown of traffic statistics. You can either view stats for the whole network of for a specific node. The stats can also be saved as a Notepad file. A principle function of the reporting tools is to create regular network health checks. You can do this at any time by loading the Report Generator, as described above, then choosing either the HTML or Spreadsheet option to generate the report from your selected stats file. The latter option creates the data within Excel then squirts it out into a Word report, so you need these two components of Microsoft Office to be installed on the PC hosting TeVISTA in order to run this version of the report. You may also find that you need to change a default file location setting within Excel the first time you run a report, but onscreen instructions are provided on the fly by TeVISTA.
The HTML report loads directly onscreen and provides a relatively simply but very clear tabular breakdown of general network stats, plus busiest users and top protocols. It then gives a variety of breakdowns of network usage in graphical, pie chart format. As such it is best used for at a glance checks and has the benefit of being readily understandable by non-techies!
For a more detailed health check report you need to opt for the Spreadsheet-Word option. This auto-generates a complete Microsoft Word document, with table of contents, pre-defined sections and integrated graphics. The idea here is to provide a �board-level� read so a significant amount of pre-formatted text is included explaining the various sections. These include overall network usage, peaks, errors, Top 10 nodes and conversations, protocol usage, packet breakdown statistics, nodes and conversation analysis and broadcast and multicast packet reporting. Usefully, the report looks to pinpoint problem areas by outlining acceptable threshold levels so you can see where they have been exceeded. It also makes suggestions in each section as to where to look for potential problems. In addition to the reports you can also upload any saved packet capture traces and print these out. There is no doubt that TeVISTA is an ambitious product. Compared to many network management products it has a huge range of integrated functions. It largely succeeds as both a one-stop-shop solution for the small to mid-sized companies and as a �take the bits you want and forget the rest� solution for the larger enterprise, or even mid-tier Service Provider who will already have several management products in place. Obviously, as with any product that covers as much ground as TeVISTA, there is room for small improvements in several areas, but equally there are no glaring weaknesses. What TeVISTA does have are two specific "uniques" � the HSRmon protocol and the Visibility Agents. More specifically it is the combination of these technologies that sets the product apart. Anyone who doubts this simply needs to see just how feasible it is to manage a network down a 9.6K GSM connection using the product. Then try doing the same with anything else that�s available out there� Finally, there is the cost justification argument � a product has to be able to add to the bottom line and help make a company more profitable. Our ROI table featured earlier in the report indicates that TeVISTA would have one of the fastest ROI turn-arounds of any product we�ve yet looked at. As such we have no hesitation in awarding TeVISTA our NSS approved rating. We suggest that not only should the SME and Enterprise level users take a serious look at the product, but a few well-known network management product vendors might like to consider OEM�ing the technology as well! Chevin even provides a �virtual� demo of the product across the Internet so check out its web site for more details.
Appendix A: How TeVISTA can reduce the number of network management components required It is common for any mid-sized - or upward - network to feature a number of different management products, deemed necessary in order to provide a complete solution. In many situations this is the case, but at the SME level the concept of providing a single product to replace maybe three or four different utilities has obvious appeal. Not only does it reduce training costs but, on a day-to-day basis, it is a far easier solution to manage. Looking at a single application, rather than loading up several different ones, has obvious benefits. Within TeVISTA�s Network Asset Manager (NAM) you can directly control many different vendors network devices and � in so doing � remove the need to use the software supplied with those products. During the testing of TeVISTA we configured the network with a wide range of devices from 3Com, Foundry Networks, Nortel Networks and Intel. In each case these products were supplied with configuration and management software which we duly loaded up. However, in each case, using TeVISTA we were able to directly manage those devices without the need to use the vendors� own software. Moreover, TeVISTA enabled us to then manage those devices in far more detail than with the tools provided by the vendors.
Without looking to make a point about a specific vendor but just to state examples, Intel�s software � DeviceView � only let us see its own switch on the network and ignored everything else. It is purely a configuration tool, but we were also able to configure the switch from TeVISTA via a Telnet session to a text-based but menu-driven utility residing on the switch. 3Com�s Transcend software at least provided a limited discovery option and some basic monitoring facilities but the configuration manager � web-based on the switch itself � was equally accessible from TeVISTA without loading any other applications, rendering the 3Com software redundant.
Where TeVISTA provided detailed mapping of the whole network, Transcend simply drew clouds with question marks! Moreover, the vendor�s own software is also less than reliable at times. For example, 3Com�s Transcend was telling us its 3300 Ethernet switch was non-contactable when a DOS ping to the same IP address proved successful!
So TeVISTA really can do the job of many different management applications - if required to perform that role in your organisation - and more reliably so than the vendors own software solution often can. Appendix B: How Network management enables real E-business and e-commerce Opportunities The �E� era is well and truly with us and � despite some spectacular .com failures recently � will be around for a long, long time yet. It can certainly be argued that global, high-bandwidth connections will offer great opportunities in e-commerce and e-business, largely limited only by an individual�s imagination and the inability to create a genuine business plan. The argument equally follows that voice and video will finally have real alternative media to operate on. So, on the back of the growth of the Internet, we have e-business and e-commerce. But is this just more hype or is there real substance and business opportunity here? Consider first that many felt exactly the same way about the Internet itself until recently and the personal computer before that. Even the telephone took the best part of a hundred years to become widely accepted. But imagine if you�d been the one to think of direct, telephone-based sales in the �50�s or �60�s in the US when suddenly it became more than just a device for speaking with friends and relatives. That there are similar opportunities now in the e-world is very clear. But how, exactly, to go about taking advantage of this situation is less so. Not least is the issue of what is the difference between e-business and e-commerce and which is applicable to yourself and your company. Then there�s the prickly subject of investment costs and � the bottom line � how do you actually increase profit through this investment and what kind of timescales dictate the return? If e-business is defined as inter-company business carried out across the Internet, then e-commerce, within the same breadth of definition, is transactional business, often an individual making some kind of personal transaction across the Internet. Classic examples are home banking, online travel reservation, or buying products, such as books and CDs, online. In all cases, e-commerce is effectively the communication between a users� PC or Mac and the remote Host computer, which is replacing either a telephone call or the trip down to the shop, bank or office, which would have been necessary before. Obviously, this concept applies equally to a staff member working on behalf of their company as to a home user ordering the latest �Best Seller� from an online bookstore. These are very strong arguments in favour of computer-based commerce or e-business. As such, they validate the near obsession with companies still throwing their services at and onto the Internet currently. But � and this is a big but � the benefits of e-commerce are lost the moment a web-site becomes unavailable or even if it is simply very slow in operation. With pro-active network management in place, however, this all too common occurrence � we experience it daily when web-surfing � can be prevented. Many are the documented cases of would-be e-commerce customers finding a target web-site down and shifting their custom to a competing service provider whose web-site was up and running. Not only that, but the chance of those customers permanently making the shift to the new supplier is extremely high. This is a market with high elasticity. So one moment�s network downtime can have a lasting effect on lost income potential. The same applies in the situation where the user finds one web-site performing too sluggishly for their patience to withstand and therefore moves on to a competing site that is performing better. While this latter scenario could be simply down to poor web-site design, equally it could be the result of a problem somewhere on the network, a problem which the IT staff need to be made aware of as quickly as possible � another role of network management. A different management perspective to e-commerce and e-business is that it can improve bottom line profit by reducing operating costs, often substantially. Offices, warehousing, staffing levels and any other fixed operating cost can all be potentially reduced by putting at least part of a business operation online. A recent study from Giga Information Group in the US suggested that the cost savings globally through business use of e-commerce will rise from what was $17billion in 1998 to $1.25trillion in 2002. As a very interesting comparison, business-to-consumer e-commerce trade over the same period is set to rise from $8billion to $108billion, so it is clear where the primary objectives of e-commerce lie in the short-term at least. But this may be for the wrong reasons � the negative attitude towards e-commerce and the obsession with its failures to date. Another argument is that, under the influence of e-commerce, services will actually become more important than that products themselves as organisations will collaborate as and when needed to provide one-stop shopping points. A commonly used example is the airline business where you will soon be able to not only book a ticket online but, at the same time, be offered a choice of hotel rooms, car rental options or other onward travel options and travel insurance, for example. Not that � using this example � an airline will need own a hotel chain or a car fleet; it will simply act as middle-man to integrate all these �product� options under one service, earning commission from the respective hotel chains, car rental companies and transport companies. What this means in practice is that the opportunities presented by e-commerce and e-business are not restricted to what one company alone can offer but are truly powerful when you consider how effective the bundling of services and products might be. In turn, this dictates the need for a truly global, high-speed network, which brings us neatly back to our starting point � the wonderful new world of near unlimited global bandwidth being touted by all and sundry currently. However, the answer lies not in purely adding ever-faster components, but in adding some intelligence to the network in the form of network management to control those components. Such is the belief � and role � of companies like Chevin to help make high-bandwidth global networking a reality. Indeed, as the speed and complexity of networks rises at a near exponential rate and the Internet takes on ever greater importance, so the role of network management becomes ever more critical to the success of e-commerce and e-business ventures. Contact:
Chevin Ltd����
|
 |
Send mail to webmaster
with questions or�
|
