Communications News Website service is offered by Highbury Business. This was the official website of Communications News. The content below is taken from 2006 archived content.
Industry veterans come together on content-addressed storage
31 May 2006
Industry pioneers have joined forces and founded a new company dedicated to providing scalable, high performance fixed content storage software, while reducing complexity, vendor lock-in and runaway costs. ...
Acquisitive AT Communications buys again
31 May 2006
AT Communications, a business-to-business systems integrator, has acquired the entire issued share capital of Britannia Telecom Group, a specialist voice and data reseller, for a total consideration of £3.84 million. ...
Rostrvm Solutions extends integration with SpliceCom maximiser
31 May 2006
Rostrvm Solutions has released enhancements to its rostrvm call centre application suite extending integration with SpliceCom's maximiser business telephone system. ...
Vodafone releases business email on new Nokia E61
30 May 2006
The E61 from Nokia is now said to be spearheading the range of compatible devices on Vodafone Business Email, and has broadened the choice of handsets compatible with Vodafone Business Email to over 25. ...
Sirenic launches mobile push email
30 May 2006
Software development company, Sirenic, has developed and launched an intelligent “push” email and data delivery service to complement existing “pull” technology, as well as introducing a number of additional enhancements to its mobile email, PIM (Personal Information Manager i.e. calendar, contacts,...
Zinwave brings together in-building mobile systems
26 May 2006
UK start-up Zinwave has announced the launch of its first product supporting its unified wireless infrastructure vision. ...
Eicon Networks, CT3 and GFI come together on fax server solution
26 May 2006
Eicon Networks, a leader in providing world class products and technologies for media processing, has announced that computer telephony distributor CT3 Europe will be bundling Eicon’s Diva Server products with the GFI FAXmaker product from messaging software provider GFI Software to...
Red-M and iBwave create in-building design tool
26 May 2006
Red-M, an independent wireless network solution and services provider, and Canada-based iBwave, a worldwide provider of in-building wireless design software, have announced a partnership to integrate the companies’ complementary flagship products. ...
Logicalis and Silicon Plains announce alliance to help customers align existing ICT systems and data with business processes and people
25 May 2006
Logicalis has signed a strategic alliance agreement with Premier IBM Business Partner, Silicon Plains, to provide its customers in the Insurance, Banking and Government sectors with an accelerated approach to greater business agility. ...
New Cybertrust intelligence service available through managed security services offerings
25 May 2006
Cybertrust has announced the addition of Cybertrust Intelligence Service to its MSS portfolio. ...
ROUND TABLES DISCUSSIONS
Gaining control of the mobile workforce
10 Nov 2005
Communications News gathered together seven leading experts from the communications industry to discover the drivers and challenges facing enterprises mobilising their workforce. Sponsored by Sirocom, the UK's fastest growing virtual network operator, the roundtable highlighted a range of critical issues. Here's what the experts had to say...
Richard Thurston: Lars, as one of the country's top analysts on enterprise mobility, how interesting do you think this area is?
Lars Vestergaard: Enterprise mobility is like Claudia Schiffer. It's very good to look at but very difficult to handle. Mobility is going to increase the number of hours IT managers need to work. Right now they are petrified. They are using security, and lack of it, as the worst possible excuse to get off the case. [But] as soon as you start working with them on the strategy their fear of security goes down. If they don't yield mobility now, they do exactly the opposite of what they want to do -- security breaches happen. You have rogue employees who activate wireless LAN on their laptops. They buy a BlackBerry.
RT: Is it fair to say the definition of mobile working covers people that could be physically within the building's walls, as well as at home or on the road?
LV: It accounts for all of it.
Lance Spencer: Mobile teleworkers are only half of the remote workers out there -- the other half are home-based.
RT: How do enterprises generally perceive mobile working?
David Macfarlane: We find two types of enterprises. Those that absolutely say no and tottle out because 'We cannot manage the security element of this' -- their policy tends to be what the clever users have delivered to them. The other side is the organisations that have looked at this objectively. They're delivering simple access to the corporate infrastructure but they're not considering the security concerns.
RT: Andrew, you're implementing a mobile workforce strategy in your part of Visa. What exactly are you rolling out?
Andrew Vorster: I am using the iPass client from Sirocom for my mobile solution. We have underutilised the current technology due to security restrictions or just general apathy within the company of trying to go wireless. [But] it's high on my agenda at the moment because it's our corporate goal by the end of this fiscal year to have 50% of our staff based in regional offices and only 50% remaining in our London office.
RT: What type of staff are mobile in your business?
AV: We have less than 20 permanent home workers. Our focus is on mobile workers -- people that travel. That has been our biggest pain point. To me, all my regional office workers are remote workers. I don't have any technology in those offices apart from switching technology. So I have the same issues with patching, security, data management, access control that I have with a laptop.
RT: So you've got all these mobile workers, and you didn't formerly have a policy for them. What business challenge were you facing with those workers?
AV: The business challenge for me is I'm after that ubiquitous working anywhere, anytime. My challenge is to find a simple solution that says to anybody this is how you access the network when you're outside Paddington.
RT: You said you were rolling out an iPass solution. How many people are on this solution?
AV: We've got about 150 on iPass altogether, but they're only using it on dial-up at the moment. They're not using broadband and they're not using WiFi. Those using broadband and WiFi is just a small pilot group of about five users.
Doug Loewe: Our next generation client will actually say 'great, the user is in this location, based on the policy set by Visa, with the most cost effective, with the most simple, with the most secure basis. But today we're reasonably pleased that we can at least give the user a choice and it's intuitive and it's easy to use.
Peter Crowcombe: The feedback from corporates is mobility is what they need to do.
RT: What are the drivers, or benefits, of mobilising the workforce?
DM: The obvious one is people wanting to retain staff. If you are an employer and you want your staff to always work in the office, you will lose your staff.
LS: Local government is a big driver. If you've got disabled children, you can insist on working from home now. In local government they are taking this up [teleworking.] And there are virtualised call centres which have employees working from home at special hours of the day.
RT: What are the inhibitors or barriers to mobile working?
AV: A lot of the time, the technology is too complex. The device must give your best choice of connectivity whether that's 3G or GPRS. The technology is just a little bit lacking to be able to make it idiot proof. User education and a lot of hand-holding is necessary until it becomes simple enough.
DL: I agree. Users are saying don't force me to choose from all these various technologies. Just connect me at the most cost effective, most secure and easy-to-use way of getting my information.
LV: It has to be plug and play, not plug and pray, for the ease of use.
DM: Also we are held back on cost. There is the perception that letting people loose with wireless hotspots [is expensive]. But there are corporate packages. The biggest challenge is education.
RT: Are there any other barriers?
LV: Mixing up work and personal life is, at the executive level, a no-brainer. The challenge is when you start moving down the organisation - the costs then need to be justified. How do you get to those people? They might not be able to stand the thought of the boss checking up on them.
DL: If it's a sales guy that thinks they're ging to be more over quota by using it, then they'll be using it more, but if it's seen as forced then workers will resist.
AV: We found it completely the other way. We had people clamouring for the mobile technology.
Rob Civil: Look at taxation. Books are zero-rated, but communication services are not. If the Government really wants to make workers productive away from their desks, then they should use taxation as a tool.
PC: There's a huge consideration. Does the management trust remote working to work? Some companies have banned working from home and want you in the office.
RT: Are there any mobile access technologies which are particularly important at the moment?
DM: SSL technology has been around for 2,3,4 years. Lots of enterprises have ben using IPSec, but in fact you should be using both. Should we deploy these cumbersome thick clients onto our laptops? Absolutely not. SSL technology is coming along and completely rewriting the rule book. Thin access, thin deployment. That's the radical change.
RT: OK, let's give it a quick vote. Yes or no -- do enterprise ICT managers have control of their mobile workforce?
LV: At all the IDC conferences, I ask how many [ICT managers] are in control of mobile working in their company. I would see one, two, five hands out of a hundred.
RT: That's a big no. Anyone else?
RC: This deperimeterisation thing is scaring people to death. You mentioned wireless. The moment you give someone a machine and say go outside the network, their reliance on good user behaviour is a real issue for people.
RT: Security is a very big consideration going forward, then?
AV: If you say 'are we in control of our mobile workforce,' yes, because their access is restricted. We've locked them down so tight up until now. We want to ensure that as we allow them to do more, that we stay in control. The policy side is important to us, but how do we slowly open that up? From our security department side, I have got to do a huge amount of PR with them.
PC: There are several things coming together -- one is security of data on the device. The other is regulation -- you have to somehow control that data, and prove you can control that data -- and the third one is cost reduction. What we're seeing is people 'webifying' access to clientless access to the centre. But the big thing that screws up the client server model is the performance you get on applications. So they are looking at how you accelerate the applications as well as giving secure access. One bit we need to solve is how to accelerate that application.
DL: It's no longer about perimeter security. It has to be pervasively secure. It might be device authentication that iPass brings or end user authentication from RSA or nCipher. You have to have one person to solve the equation.
RT: OK, what about cost and billing? You have fixed, mobile, 3G bills. Is that an important issue for ICT managers?
DM: We tend to find that the enterprise doesn't look at the whole gambit of cost that comes from a mobile solution. Companies often expect mobile workers to put their WiFi charges through the expenses route. That's a hidden charge.
RT: Do you know how much you're spending, Andrew?
AV: The majority of the cost has been through iPass. But we have to educate users responsibly, that if they connect to a WiFi access point they have to remember to disconnect. But the reason I don't know in its entirety is that at certain hotels you have to pay for access on a voucher basis and that does go down the expense route.
DL: The pricing for a combination 3G, 2.5G, WiFi, DSL can be a flat rate per month.
LV: It costs an arm and a leg. One megabyte of roaming on 3G at the moment costs about 5euros.
RT: So a big cost issue there. Is there a big time issue as well?
DM: There certainly shouldn't be if you bring it together in a simple interface.
RT: Should ICT managers have to prove a return on investment argument for their mobility plans? Andrew, did you do an ROI?
AV: No, because we didn't do it to reduce desks. We did it as 'we want to get our people in front of customers'. So I'm not even trying to justify what I'm doing. We've just rolled out BlackBerry. The cfo wanted an ROI case. I said 'I can give you an ROI case based on the extra productivity. But if you want ROI for a BlackBerry, can I see your ROI for a fixed line telephone on someone's desk?'
LV: My recommendation for a vendor is have an ROI-type calculation for the cfo, but ROI is not the main driver.
LS: One of the big inhibitors is getting that ROI model put together. The costs are distributed so widely in the organisation, it's hard to calculate. For most companies there isn't a straight driver you can take to the cfo and say 'here's the model, here's the payback'.
RT: Any final thoughts?
DL: Connecting your remote workers is no longer a case of can we afford to do this, it's like oxygen -- you have to have it. The real question is how do we go about doing this? Ease of use, security and it needs to be cost-effective. It's no longer a matter of if we're going to do it, it's how and when. It's about opening the floodgates and allowing workers to be productive.
AV: I've got more confidence now that there are a whole bunch of people working on this. I keep on thinking 'why can't I make this simple enough for my president'. I haven't lost hope.
LV: I'm not wrong about knowing mobile working is going to happen, I may only be slightly wrong with the timing. Email is now, right? And then it's just a question of how quickly you roll out the rest of it. Everything is happening now -- taking those processes or steps to mobilise. Five minutes or five years, I don't think we're very far from there right now.
The great virtual network operator debate
19 Jul 2005
Businesses have traditionally bought their networks from a single supplier, typically BT or an alternative carrier. Some organisations have built their network themselves. But now a new breed of network supplier is emerging -- these are virtual network operators (VNOs) which can pull together the best suppliers for each individual part of the company's network. This roundtable, hosted by Communications News and sponsored by Sirocom, asks why VNOs have become successful, and the advantages and disadvantages to businesses of opting for a VNO approach.
Richard Thurston: Businesses have a wide array of choices when procuring their connectivity. What choices are they facing?
Chris Lewis: There is a continuum from the total DIY solution, where you design it and purchase it all yourself, all the way to a fully outsourced solution. But what I find most interesting is not the extremes, but the middle ground, which is probably 80-90% of the market, which is where customers keep some things in- house and source some other things from a third party.
RT: OK, thanks for outlining those options, Chris. What should businesses be asking for from the supplier of those options?
Wayne Churchill: Its application performance.
David Macfarlane: Users are still buying networks on old metrics -- roundtrip packet delay, loss, jitter and availability. What we need to do now is to say is we can guarantee SAP [or whichever application] between Site A and Site B.
Roger Chadderton: The first thing we looked for was stability: how long our supplier was going to be around. Then we needed to establish the breadth of services and the relationships that they had with other players in the marketplace.
Doug Loewe: Customers want flexibility, they want us to arrange a network of partners who can bring different components. It's worth embracing the idea that there are experts in certain areas of the OSS stack or in certain geographies.
DM: We recently did a survey across our customer base, and the ability to have technical and commercial flexibility within your agreement was number 1. Number 2 was customer service. The price was third, ironically.
CL: That's because price is negotiable. Price was at the top of people list under the monopoly -- it's actually drifted down the scale -- whereas customer service and technical competence are not negotiable. It's not that price isn't an issue.
RC: Yes, price is flexible, but it's not the final decision maker. Technical ability is paramount. Technical and commercial flexibility you need to get from doing reference visits.
RT: There has been a lot of interest recently in companies that pull together a solution from a range of suppliers on a bespoke basis for businesses. These companies are described as virtual network operators. David, you run a VNO. Could you explain what a VNO does differently from other network suppliers?
DM: In a pure sense, we are a carrier which owns no network. We use best-of-breed assets to deliver the enterprise the best end-to-end service you can buy by service, price performance, reach or technology. We believe flexibility is the key. Enterprises have been let down by our industry tying them down into cumbersome fixed contracts.
WC: One of the criticisms that was always levelled at a VNO is "How can you manage the network to the same degree as the original provider of the network?" and "How can you do it on an enormous geography?" The answer to that is we don't try -- the VNO is a business that uses the capability of the carriers to manage their own networks. We buy a service which comes to us with a service level agreement. To provide a best-of-breed solution you have to integrate; you can't simply resell a BT service or MCI service to a client. VNOs also give freedom and flexibility. They don't have to align their future technology strategy with the PTT's technical roadmap. If you write a 5 year agreement with the customer, you lock that customer in to your roadmap. The second reason customers would go to a VNO is to access the lowest lifetime cost. It's inconceivable that every carrier can be competitive in every geography on the planet. In given regions, there will be lower cost providers.
DM: The value-add is a number of components -- the integration of the best network for the best job in the best region. Show me a truly global operator. Also the value we add as a VNO is the customer service -- that is second to none, it has to be.
DL: What the VNO model brings to the enterprise is the diversification of risk.
RT: Is BT Global Services a VNO in a way?
Huw Watkins: In some respects, yes. We've never had a situation where everything end-to-end is BT's own assets. We will take the best of the broadband providers around the globe and we will work with different WiFi suppliers.
WC: A carrier can provide those services, but here it is Mr Customer, and here's the SLA. But a VNO has no interest in selling bandwidth.
HW: What you describe as a carrier is only a part of BT -- BTWholesale. At BT Global Services, we provide that service edge. We have the capability to buy outside BT to provide the right service for customers. Our roadmap will become much more flexible than it has in the past. We will not be tying customers into technology in the way we have done in the past. In a sense, we are creating our own VNO model.
CL: There are various models in the way VNOs can operate. Some, like Sirocom and Vanco are exactly the same, you purchase services specifically for each contract. In someone like Virtela in the United States, they actually have peering points with many of the leading providers. They choose on a day-to-day basis which network they put the service across to get the best performance. The biggest single distinction [between a VNO and a carrier] is that the VNO is buying the service for a particular customer. BT or the regular incumbent players are constructing a network which they then sell to all the customers they deal with.
HW: I don't see what we're doing is different to what VNOs are doing now. You choose the best supplier for your customers. We do exactly the same thing.
WC: It's very different. You're building a network and you make a big investment. Those are the assets you sell. If the customer doesn't want to buy that asset, tough luck. The VNO doesn't have an asset.
RT: Has the geographic scope of either carriers or VNOs changed? There has been considerable consolidation in the marketplace.
CL:We've been through this phase in the industry of saying we're going to go global and everyone pretty much has contracted back to domestic basics. BT is now in an exceptional situation, having acquired Infonet and Radianz, in that they can build more of a global presence. Relatively few people have anything like the assets in the ground to deliver a global service. If customers want the flexibility, they will buy from a Vanco or Sirocom or whoever. If they have a relationship with a BT or a T-Systems or an AT&T, they will probably stay with that relationship. It goes back to this trust issue.
RT: Are there any weaknesses in the VNO model?
HW: It's not a weakness, it's a challenge. BT has, what you describe as, wanted to sell you Megastream or Frame, whether you like it or not. That's an older view of BT which was certainly true. But isn't true any more. We have learnt that service agility will keep the business. We offer much more flexible commercial terms so that there are benchmark clauses and get-out clauses. We will be playing more and more in the VNO space. There won't be a fixed roadmap which we're going to drag customers kicking and screaming along. I think what the VNOs will have to do is to respond to that challenge. How do they respond to BT and others coming along and offering a very similar offering?
Lance Spencer: Customers should be looking to see that [their supplier] has got the skills base for what they want to do, and that they have the interconnections for systems and processes, because at the application level you have to be able to behave in a flexible manner for the end customer. VNOs can do that because they are smaller, more bespoke companies. Carriers are good at turning about mass products but have difficulty handling bespoke solutions.
CL: I think you'll find that a lot of people that have moved to virtual network operators like the transparency. I think the systems that the VNOs have developed is their asset. The only question that sits against the VNO model is how does it scale? There is so much personalisation in the way the service is delivered. And can VNOs compete with the BTs, AT&Ts and MCIs, while at the same time those carriers are taking on board a lot of things the VNOs are doing?
WC: If you look at the VNO model in absolute purity to say you offer a global VNO where you can choose any carrier in any technology in any geography, you do end up with a scale problem. So there is a limit to what a VNO can offer. If a customer wants a network built on Lucent equipment, he can't have it because I don't have that capability.
CL: The question is to what degree different users are prepared to pay for the management wrap which the VNOs provide? It's the transparency. With T-Systems the big multinational customers have no visibility of what the underlying [infrastructure] piece or the management piece is costing them. From the CIO point of view that's frustrating. That's the incumbents challenge -- to say we're not ripping you off on the underlying piece, we're offering you a managed service.
HW: It would be interesting to speak to any customers who feel they are being ripped off at the moment because we're competing in the same marketplace. We benchmark.
CL: The benchmarking exercise is a really difficult one. A lot of contracts are written so badly that being able to properly benchmark what those services cost is very very difficult. That is why some customers like the transparency of a VNO. If a VNO swaps out a technology and saves money, they often share that saving with the customer. Customers love that. A lot of end users don't have time to check they're getting the right service.
RT: Roger, you chose Sirocom as your supplier. Why did you choose a VNO?
RC: The point of dealing with a VNO is to take the pain of dealing with carriers away from me. We recognised that the business we would give to an MCI or a BT would be nothing. We would not be number one on their list. But now we have a win-win situation.
RT: What did Sirocom help you with?
RC: OK, we were forced into a situation where we had to change the network by a company that was selling us. VNOs were then a new concept. We went through the tender process with Sirocom and a couple of the major carriers. On cost, commercial flexibility, and customer service in my definition, Sirocom proved to be streets ahead. They originally put in a leased line and ISDN line to 13 sites in Europe and later changed that to a leased line and an ADSL line, the leased line for critical traffic, the ADSL for non-critical traffic.
RT: How much of that does Sirocom provide?
RC: They manage everything from router to router.
RT: What do you think of the trust issue that Chris raised. Is trusting a supplier a big issue for businesses?
RC: Yes it is, but you have no choice. The only way round it is to talk to other customers. But we have SLAs of a wide nature. It's not in the interests of any supplier not to meet their SLAs.
RT: The option we haven't talked about is building the network internally. How might do about doing that?
CL: The DIY approach? You can do it, but you have to do it with a lot of different people. The fundamental problem is the resource that the CIO has at his disposal to build the infrastructure. If you take Ford, for example, think of all the people that you would have to speak to for all its dealerships.
RT: Maybe it's a proposition for smaller businesses with just a few sites.
WC: That sort of customer is unlikely to have the skills.
DM: We see the DIY approach in the mid tier that do have a small networking team. But they're at a disadvantage before they start because they have no scale of spend.
LS: If you're buying broadband, it's very localised and it isn't very straightforward.
WC: In the mid market, they don't really DIY like a VNO does. The only good examples of DIY, like Shell, are like VNOs. Shell has a network management department of 350 people and they do exactly what we do. They have the scale -- 27000 endpoints.
CL: But interestingly they use Cable & Wireless to manage all the contracts for them.
RT: The reason I ask the question is because carriers like Geo just sell dark fibre to companies -- that is their business. The idea is companies can come along, buy dark fibre and put the services on it themselves.
LS: And some will, especially in the core of the network. But this is only the really big enterprises. As soon as you drop below the big 100 or 200 and into the midmarket, they don't want to do all of that. That is why the VNOs are taking off -- because the midmarket is moving off frame relay and understands the cost benefits of moving to VPNs. Those firms are really going to VNOs to provide that service.
DM: Who is going to guarantee that packet from A to B in a DIY solution? It's a very risky strategy.
CL: I think the opportunity is shrinking in the pure DIY sense.
RT: When it comes to security, do either the carriers or VNOs secure the network better than the other?
HW: VNOs will have to look at how they provide end-to-end security. Because we run the network, we can see security threats. What the VNOs are doing is taking security and response information from BT and other suppliers.
WC: That's a wholly unconvincing argument. I don't think there's a case to say that because you've got the physical asset, you can secure the data any better. You're providing me exactly the same as you provide to the end user.
LS: VNOs, because of what they do, almost specialise in security.
DM: Because of the visibility we have of the network, we can even empower the carrier to say where the faults are. And I can see that someone is running Skype in port 80 and that it's not http. We will tell the customer that. That helps the security.
LS: Management like that matters once you start running voice over the data network.
RT: Any final thoughts?
DM: No one communications carrier can meet the complete communication needs of the enterprise. As a VNO we are here to challenge and change the way organisations acquire communication services. The VNO is now recognised, it's being embraced more and more by the enterprise. It's a very compelling model.
LS: We see the model of distribution in the UK being broken down from one-PTO-serves-all to a number of layers of distribution.
CL: The choices that end users have are increasing. It comes down to what level a user wants. If they don't trust a third party, they will do it all themselves. Then there are all these gradations, combining more levels of managed services, be it security or the VPN or whatever it may be. The VNOs have changed a lot of people's thinking.
WC: There is a sizeable shift in the attitudes of enterprise in terms of how they acquire networks. I don't think we have to persuade customers about the merits of the VNO context these days. I think there will be customers who will also stay with the incumbent carrier -- those organisations that are slightly more conservative and less innovative. But if a customer wants freedom of choice and flexibility of technology with the lowest lifetime cost and the highest service levels, then they'll contract with a VNO.
Solving the bandwidth dilemma
24 May 2004
Participants : Mark Logan (BT), Scott Dobson (Expand Networks), Nigel Thurlow (MacFarlane Group), Tony Reeve (Omnetica), Sean Fane (Interquad Systems) & Vince Potter (DV02)
A Communications News roundtable in association with Expand Networks
A Communications News roundtable in association with Expand Networks
New application rollouts place a big strain on the wide area network, and can often be cancelled due to increased bandwidth, and network element, costs. Communications News and roundtable sponsor Expand Networks assembled industry experts from across the supply chain to discuss possible solutions to this pressing problem.
Keith Dyer: Scott, could you outline for us what the bandwidth dilemma is?
Scott Dobson: Organisations are looking to deploy applications and services across their wide area network and those applications and services tend to be quite bandwidth hungry. I see three camps of users. There's the optimist network manager who delivers the application and hopes the network infrastructure will support it. Then there's the pessimist, who does his due diligence so if something goes wrong he has a contingency plan. Then there's a third group of people who are not deploying applications because they know their network can't support the application. Defeatist, I suppose you could call them.
KD: What options do these three groups have?
SD: The obvious one, which happens 99.9% of the time, is to upgrade the telco circuits -- to give more bandwidth to the branch network and deliver the application. But you're paying for that on an ongoing basis and lead times can be quite painful. Waiting 30 or 60 days when the application is already deployed is painful. Another way of optimising the network is web caching, or proxy server or content delivery, which is part of Cisco's portfolio. This is about delivering web content closer to the user, which is a good solution if the applications are being run from within
the browser. Gartner's analysis says that between 70-80% of enterprises' wide area applications are non web-based
-- they're native applications like SAP, Siebel, Exchange etc.
Another solution is bandwidth management, which you can buy in appliance form from the likes of Packeteer, Sitara and Allot.
You can also buy next generation IP VPNs with MPLS quality of service in the network itself. A great solution, but if you need more capacity, you need more capacity. What bandwidth management allows you to do is to prioritise critical applications. And lastly, data compression. It's been around for donkey's years with zip and mpeg. If the data is not text-based, data compression is not very efficient. Those are the options.
KD: You mentioned three types of network manager. Do you really think there are that many defeatists?
SD: There are application rollouts, whether it be Citrix or Siebel or SAP or Oracle through Citrix that get put on hold because when they investigate how much it is going to cost to build the infrastructure to do it, it just loses the compelling reason to do it in the first place.
Vince Potter: We have a customer that has an application at the moment that has cost them millions of pounds that they cannot roll out at the moment because their existing infrastructure will fall down instantly as soon as they deploy it. We have shown them Expand, and they realise they have to do something with the network. They know they don't have to throw bandwidth at it.
SD: Ignorance is ours, and our competitors', biggest challenge. People don't know there is an alternative to doing the obvious things.
KD: What sort of technologies does BT see as successful in this area?
Mark Logan [via conference]: Class of service [MPLS] with some caching technology and something like Packeteer-- combinations like that work well.
KD: Scott, one of your biggest deployments was in the financial sector, wasn't it?
SD: Yes, Abbey National deployed our solution on 755 branches across the UK. Siebel was the driver, that was the only thing they were interested in delivering. They had to ensure they could deliver that solution because it cost them x million to do it. So our solution, which came in at £3million, was inconsequential compared to the importance of that application.
KD: Nigel, you implemented WAN optimisation. Tell us what you did.
Nigel Thurlow: MacFarlane, for the last three years, had been operating on Cable & Wireless frame relay technology. They were originally a very distributed network using distributed computing architecture. At one point it had 46 sites in the UK. They then acquired the National Packaging Group and acquired some additional sites with Kilostream end-to-end circuits and distributed computing. But National Packaging had already started a project to go to Citrix MetaFrame and to become a server-based computing house. MacFarlane wanted to continue to move to server-based computing and make all their architecture Citrix-based, but there were some limitations on the network because they were operating 64K circuits. At the same time they were having some service issues with Cable & Wireless. When I came in, we were looking at alternatives to using frame relay due to cost, and alternatives to end-to-end circuits. We started exploring MPLS and IP VPN as an any-to-any solution. We were looking at reducing the bandwidth requirements we had but at the same time deploying Citrix MetaFrame to all the sites. MacFarlane have gone through a restructuring exercise and were operating out of 22 sites within the UK, two sites in Dublin, one in Sweden, a couple in California and one in Budapest, plus a recent acquisition in Mexico. So we needed a partner who could deploy a global network. The contract was awarded to BT Global Services.
KD: Which technologies did you go for?
NT: One of my staff had already had some exposure to the Expand product set. That brought caching into the discussion. We looked at Cisco's similar products. They were quickly discounted because of the original cost. Although MacFarlane's LAN and WAN was all Cisco-based, the cost of the Cisco caching engines were substantially higher than the Expand technology. DV02 lent us some equipment for a two month evaluation and some guys at Expand lent us some technical knowledge. We did some proof of concept on some of the existing circuits to see if the caching engines would deliver the goods.
KD: Did they deliver the goods?
NT: Yes, we are now able to run up to 25 users over Citrix on a 128K pipe and up to 55 users on a 256K pipe. It doesn't take a great mathematician to work out how big the pipes would have needed to have been had we not had the caching technology employed.
KD: Was there a resultant financial saving?
NT: By moving to MPLS, we saved £45,000 to £50,000 per annum on the cost of the network and by moving with the Expand technology we have now been able to deploy the thin client solution to all the other sites, again recognising savings. To give an example, we [originally] had to run a circuit out to Budapest on frame relay at £24,000 a year. We now pay £6,900 with exactly the same circuit speed. In the UK it was less dramatic. We calculated our ROI at between 3 and 6 months on the total spend on Expand. And there is the impact on productivity. We put a test unit in the Exeter office, and the utilisation of the bandwidth dropped from 140% of available bandwidth to 40%.
KD: Why didn't you settle for a Cisco-based solution if your network is based almost entirely on Cisco kit?
NT: I'm a big advocate of Cisco products and I've worked on many in previous jobs. But we didn't spend a great deal of effort on them this time because there was no point paying Cisco another £1000 per unit to deliver exactly the same functionality.
KD: BT helped implement the solution, even though it limited your data consumption. How did its team embrace the concept?
NT: The BT project team wasn't aware of the technology. We had to introduce them to it. And they were only just getting to grips with the MPLS technology, which they refer to as IPClear. BT had to understand the issues because they were to take over the management of the existing stock of routers.
ML: You have to remember that BT is a rather big company, and BTGlobal Services is split into two units -- a high volume low unit-cost business and a solutions team which works on a customer-by-customer basis. The solutions team would know more about Expand.
KD: Apart from educational issues, did the rollout go smoothly?
NT: There are some foibles in this technology. You can't have any proprietary compression employed post acceleration -- we found that caused some problems. There are also some other issues. We had to have some discussions with Citrix because we had some issues in disabling compression and encryption. And we also had involvement from Wyse UK. The Wyse firmware on their terminals, even though the check box was there, wouldn't disable compression. So we had to have a beta release of firmware from Wyse. We did also have some difficulties with the initial configuration. Some of that was down to BT configuration on the IP cloud.
We got through that. It was painful and took a lot of work with BT. We took it back into the labs and went through a full evaluation.
But we got through all the problems.
KD: How did you find BT's attitude during those problems?
NT: Very frustrated. They didn't see it as their problem. However, I said, "we want to deploy this technology over MPLS and we want to deploy Expand caching". They accepted the contract on that basis. BT, to their credit, after a lot of debate did find the right people to put the right solution in place. It was amusing, because later in the project we discovered about the Abbey National contract with BT for MPLS with the Expand Accelerators. This wasn't known to their account management team at the time. We told BT, and a lot of the problems got sorted out.
BT was very dubious about whether it [the Accelerators] would work or not.
KD: Were you confident the caching solution would work
NT: We did the proof of concept and a lot of tests.
KD: During testing, did you find compression was more effective for certain traffic types?
NT: Citrix is very dramatic from an acceleration point of view because there is a lot of static screens and imagery. All you are doing is passing mouse clicks and keyboard strokes and small screen updates. But for something like POP3, which is changing quite dramatically, there is less effect.
VP: What sort of improvement did you get on Citrix? We reckon 300-400%.
NT: We had one site which was almost touching 600%. 300-400% would be a good example.
KD: Were there any risks in taking the Expand solution?
NT: The only risk we mainly perceived was failure of a unit. The only criticism is that the warranty support at the moment is a replacement on a 24 hour basis. The risk is that if you don't keep a spare unit on location and a unit goes down, the network stays up, but you have to reduce your number of users down while you don't have a unit.
Tony Reeve: Are there any management issues?
NT: No, fire and forget.
KD: As a distributor, why did you take on the Expand product, Sean?
Sean Fane: Delivery of applications is a big driver for enterprises now. They have desktops to upgrade, they have got new applications to bring on board. Operating a server-based computing network makes it incredibly efficient to send completely new applications to thousands of desktops quickly. But of course you can only run a server-based computing environment efficiently in a WAN if you have a tool which maximises the efficiency of your bandwidth. Expand was the best of the one or two products available at the time.
KD: And, as an integrator which works with a range of different companies, where do you see the sweet spot for this type of solution, Tony?
TR: Maybe the larger organisations are more educated in these issues. [But] there are probably many smaller businesses that are rolling out applications without really considering what it will mean to their infrastructure.
KD: Can telcos leverage these technologies?
VP: If you take the global providers, I think it's a massive opportunity for them to enhance their services. The likes of Equant, Infonet, where they are delivering a managed service.
TR: There was the view of the 'One Network', where the service provider was going to be the IT utility provider and was going to push itself out to ownership of the end router and the WAN.
SF: That will come back with a force. But I think it is more a case of will larger enterprises take that leap of faith in outsourcing their applications? It should be the telcos, in partnership with the systems integrators and specialist resellers.
TR: Yes, because the telcos don't have the skills to deliver on their own to the end user.
SD: The telco sells bandwidth, that's what it does. That's where it makes its margins. Why should it sell something that's better for the customer, where it doesn't make as much profit? Until they start offering these solutions, they really need to decide whether they will offer customers best-of-breed solutions for their problems or whether they are a telco selling circuits.
ML: I think deploying edge technology like this can work for the telco. It allows the user to free up bandwidth so they can put another application, such as voice or video, down that pipe. It's about helping customers, not cannibalising revenues. Network managers talk about using these technologies for 'freeing up' bandwidth so they can use it for something else, not for a network downgrade.
KD: Does anyone else think edge optimisation will help drive voice over IP usage?
SD: I talked about optimist, pessimist, defeatist. Voice over IP has a globe full of defeatists. People are just not deploying it across their network because they know they haven't got the infrastructure.
NT: You have to reserve so much bandwidth for voice and you've got to ask what's the cost of that bandwidth. Anything that reduces your data bandwidth leaves you more space for voice.
SD: One company that springs to mind is a company with a 128K link from Dubai to London, which is already accelerating data. They came to us to ask the implications of deploying voice across this accelerated link.
As long as we can view the application, which voice is, and its IP-based, then we can prioritise it. We can't compress the voice, because it is already compressed.
NT: There is one other problem. Take server-based computing and take voice. Put them both down the same pipe. They are both real-time protocols. Which is going to get priority? Then we run into the argument, do we need to channelise the bandwidth, and that makes the bandwidth even more expensive.
TR: From the point of view of IP telephony, bandwidth is really an issue that is holding back a lot of customers. There is a good opportunity to marry up a solution like this with a voice over IP solution.
KD: If you had more bandwidth than you knew what to do with,
Nigel, would you deploy voice over IP?
NT: Yes, I would. I would do it tomorrow, if the bandwidth was available at the right price.
KD: With your current bandwidth, would edge acceleration combined with MPLS preserve you enough headroom to run voice over IP?
NT: No. If you start to look at voice, every channel needs about 16K. You are probably going to have to add 128K because that's how MPLS works.
SD: Keith said that sooner or later, you will need more bandwidth. But if you can bring a return on investment until you need that additional bandwidth, you're saving money -- that's really the compelling argument.
NT: But the extra you'll need to buy is not as much as the extra you'd of needed to buy if you didn't have the Expand technology in the first place.