CIO's Perspective: Managing 'change' within the enterprise
Date: Wed, 07/04/2012 - 18:17
Networks are becoming ever more complex to manage with the growing need to monitor and optimise the performance of an ever increasing array of individual business critical applications in addition to broader network requirements such as maximising throughput
CIOs are faced with the choice of managing this complexity in-house or using the capabilities of a managed service provider. To obtain the highest levels of performance should CIOs trust a third party to manage their networks or can they gain greater benefits by using best of breed technology vendors and managing in-house. Peter Hall, Principal Analyst at Ovum, will put the alternative approaches to a panel of leading technology vendors at NetEvents EMEA Press Summit, Garmish, Germany, who will argue the case for which approach is best. In addition the CIO has a growing role in the organisation of being a champion for innovation and change and ensuing that the benefits of technological change are realised as an enabler for new business models and ways of competing. How do technology vendors assist CIOs in this important role.
Panellists: Charles Ferland, BNT Vice President, IBM System Networking Europe; Carolyn Crandall, VP Global Marketing, Riverbed Technology; Daniel Beazer, Senior Research Analyst, EMEA, Tier1 Research; Brett Johnson, VP Business Development, AboveNet Communications
Introduced and Chaired by: Peter Hall, Principal Analyst, Ovum
Welcome everyone. I am really pleased to be here. I am Peter Hall from Ovum, based in the UK. I do not know about you, but I thought that last session was absolutely fascinating and learned a lot from it. I think we have a tall order to match that, but this topic is very broad.
Managing change within the enterprise. So I thought a bit about how to tackle this, and I had originally put a whole bunch of charts together. But what I really want to do, I think, is give the panel plenty of opportunity to discuss their own perspective of it.
We have two vendors, one service provider and one analyst. That is a great mix. I am really going to focus on one chart just to introduce that. If you get this at the end, there are a few others at the end of the facts.
The CIO role these days is unbelievably complex, as we heard that from Dan during his session earlier this morning. If we look at some of the things that are facing them on a day-to-day basis, of course for most large enterprises at any time there are really significant strategic transformation projects which are life-changing for the business.
They might be enabling new channels and bringing in new customers through different media and so on. Those things are things that have the visibility of the CEO and the board and are always going to be pretty close to the top of any CIO’s agenda.
You can always classify the role as keeping the lights on and things that are transformational like that. A lot of the things that are keeping the lights on of course are also ensuring that mission critical applications that the enterprise is running are running to a high level of performance and if they do not then that has a major impact on the company’s bottom line. Of course, if we look at the networking area, the challenges there are always getting greater in terms of the demands of applications and over the last five years real-time apps and video, the rise of telepresence. They are really pushing up the requirements there, in terms of networking, in addition to all the mission critical apps. A large enterprise is probably running a hundred or more applications, most of those mission critical.
The number of devices and endpoints is increasing as well. We have only really seen the tip of the iceberg if you think the impact of machine-to-machine has really not hit enterprises yet. That is also going to have a colossal impact as the number of endpoints being connected and being managed in the enterprise, which goes up potentially by orders of magnitude – really dramatic increases.
Then there are things like bring your own device as well which are almost things that are just irritation factors, but nevertheless it does reflect a change in the way employees want to operate within businesses, which CIOs cannot ignore. It is usually the board members who bring these devices first and they have no choice but to ensure that they can be connected securely. The whole security issue goes without saying, and I know the next session is covering that.
We hear all this hype about cloud computing, but cloud computing of course is going to have a huge impact on enterprises, but probably is not on the top of CIOs’ agendas today. They will be doing projects that fit alongside a lot of other activities in looking at new sourcing strategies, whether they are moving more to managed services, whether they are moving more to outsourcing and so on.
Geographies. If you are a large enterprise, then a lot of the attention in terms of networking is going to be bringing new sites on board globally, and there are plenty of multinationals these days who require connectivity to upwards of 200 different countries, which itself is a pretty big challenge.
Demands from lines of business. I think this is probably going to have a profound effect over the next few years because in many enterprises now, with the sales and marketing functions in particular and sometimes others, are really starting to exercise their muscles in wanting to procure some of this IT themselves. They can potentially do that with Cloud computing and other technologies? Does the CIO just try and resist that? Probably not, they probably work with those divisions so that they are not necessarily in total control but at least they are aware of the things that are happening.
This has been happening for a while of course. Organisations taking sales course for instance, for many that is procured by the sales department with the understanding of the IT department but little more.
A lot of CIOs I am talking to now are finding this role of becoming a champion for innovation, this is an expectation the CIO now has, particularly from the board. They are keeping the company up to date with not just the technology advances of IT but other advances in terms of the impact on the business. These might be coming from all sorts of areas in terms of scientific and engineering advance.
So you can imagine that the role of the CIO is pretty complex and balancing all of these different requirements is a tough job to say the least.
That is all I am going to do by way of introduction. We have a panel of four companies and I think this is a good time for them to come up now.
Charles Ferland, BNT Vice President, IBM System Networking Europe
Carolyn Crandall, VP Global Marketing, Riverbed Technology
Daniel Beazer, Senior Research Analyst, EMEA, Tier1 Research
Brett Johnson, VP Business Development, AboveNet Communications
I would like to start by taking this from a very high level general viewpoint. I would like the panel to make a comment on how their company views the key issues of change within the enterprise today and how their company is able to assist enterprises in addressing this. We will go from right to left, starting with Charles from IBM.
Clearly IBM is a huge organisation with a massive role in systems integration and outsourcing and all of those things. I know you are representing one part of it but could you try and give us the big picture initially in terms of what you see are the major areas of change that IBM is being asked to address today?
From a networking perspective, we are looking at changing it as an evolution more that a revolution in some way. As we see OpenFlow and software defined networks being important, it certainly has its future in the network of the future. It is not just about adding more speed and more technology, it is making it smarter not just faster, it is about being a smarter network having a smarter infrastructure. This is where we are looking at deploying this new technology more in an integrated and optimised system rather than just having a breakthrough technology deployment across all the infrastructure and saying that tomorrow or next Monday this is going to be an OpenFlow infrastructure or this is going to be a software defined network. We are looking at integrating part of the network, integrating part of the system into more optimised for the application for the business needs and having that deployed and connected to the, I want to say to say legacy infrastructure, but what is today’s current infrastructure.
Are these demands coming to your primarily direct from CIOs today? Where are getting this? Where are feeling this demand coming from?
More from strategic point of view, more from the CIO than the operational people. I think that from a tactical point of view, people are happy or comfortable with managing the infrastructure that they have. As we are going up into a more strategic or holistic point of view, we realise that we cannot throw more band width at the problem and hope that it will go away. We need to do something different. The network somehow did not evolve as quickly as the application, and I think it was Dan Pitt this morning mentioned the [servicing] infrastructure, the application and the storage infrastructure all being virtualised, and we are really just at the beginning of network virtualisation. From a strategic point of view, I think that more our customers are asking for more integrated and more flexible systems than the operational people.
Clearly, this is a networking conference. We are going to major on the networking aspects of this change during this discussion, but we will also be coming on in a moment to talk about some of the implications for data centres as well.
Brett, can you just give the service provider perspective on how you see this change impacting your business?
Sure. Just so that I am able to give the context of where we are seeing it from, we focus on high width band services in major US and European markets and we connect all those markets together. What we are seeing is not just the data being stored remotely, but applications being moved to remote locations. That converges with the other phenomenon we are seeing. We are seeing data centres being moved further out so suddenly the network becomes more important in terms of both high band and low latency. Three of four years ago, our customers were never asking about latency at all.
Today it is one of the first questions they ask us. What is the latency between these locations because when you are running mission critical applications, latency becomes very important, particularly machine to machine type applications. These applications are no longer sitting in the server room, nor are they sitting in a local data centre.
They are sitting in data centres in different cities or globally. So that is one of the big trends that we are seeing. I think it is a bit misleading to call it Cloud computing in that it is a phenomenon that has been going on for a lot longer than that term has been used.
Carolyn, your business is around ensuring that applications’ performance is optimised and issues like latency are also things that you are key to addressing. How do you see this impacting the enterprise? At a high level and we will drill down to some of the other aspects in a moment.
As people look at all of the changes happening, whether it is with consolidation, virtualisation, Cloud, whether it be public private, everybody is being faced with new performance issues. When we look at performance it is from the view of the user, whether it be the actual employee or the consumer coming in to experience something on a website. What does that experience feel like? How does it feel for the IT person who has to manage all of this chaos, especially when it is something they cannot control in a public Cloud environment? They are used to having two devices that they can control and now they cannot control the [mood] changes or management at the other end. Then the third aspect of it is the financial pace. Some of the things that are coming out today require organisations to complete re-architect the way that they have traditionally approached networking.
So is that change for the sake of change or is it change for the sake of improving the experience or improving the cost savings.
Riverbed’s approach to it and one of the things that has worked well with the CIO is that we are enabling them to make those decisions based upon what is right for their business versus what the limitations of the technologies are. We can offer them the ability to analyse what those changes are before they are made. We are able to accelerate it, regardless of whether it is their infrastructure or otherwise. From an IT perspective, we will allow them to control it in a more controllable way. Just keeping it at a high level, that is really what we try to do. We allow people to go anywhere, any place, any device and still get the performance that everybody seeks, whether you are a user, IT or a business person.
We will come back to some of those things in a moment.
Daniel, your background is more with the data centre environment so . . .
That is right. I think maybe the value I can have here is perhaps taking a step back here and putting this all into context. There is so much talk of change, challenge, gosh the world is falling apart, it is coming together, etc., etc. It may be worth taking a step back and looking at what is really going on.
A good starting point for that is just what Dan Pitt said. None of these technologies we are seeing here now are new. What we are seeing is the process by which goods, services and products that were a while ago only available to those with massive budgets, becoming cheap, [victus] and widely available. I know we are in Germany.
So, as we are in Germany and I like to think of our German people and German thinkers in [Schumpeter], I hope I am pronouncing that correctly, wrote a great deal about this process of creative destruction. That is also a good think to talk about in front of an audience of people that are in the Telco world because there is no-one who knows how to have a race to the bottom like a Telco. As you can see if you look at IP transit prices over the last 10 years, this has probably created destruction of that creativity. So what we are really seeing now is, again as Dan Pitt said, in the old days if you wanted to run some software you went to IBM and they gave it to you a year later. Now you can walk into a shop in the high street, you can pick up a Smartphone, you do not even have to pay for it preloaded with applications and 64 gig of memory.
[indiscernible] but that is what you can do.
I think it is worth emphasising that this is not a new process. It is actually one of the features of what people call capitalism, although it is really not capitalism in the Marxist sense. It is really a feature of the exchange of goods and services. If you look at Wedgwood plates, something I imagine everyone in this audience is familiar with, what Wedgwood did is that they took something that only the rich had, tableware, and they made it so cheap that it could be bought by most low income families. That is what we are seeing now.
In my view this has three consequences for the CIO up there. First of all is, Peter has been talking about bring your own device, bring your own BYOD [daughter] device, whatever it is into the office. It is not just that, it is that employees can stick together a pretty decent productivity suite on their own and not paying anything, using tools like Dropbox and Google, etc. As Peter said, at Tier1 we cover the Cloud market.
We have an entire segment of the Cloud market we call the [indiscernible] Cloud market. That is just people who can create a road test and [indiscernible] environment.
These [indiscernible] the CIO know about. They have just taken their credit up and they have fired up some instances on Amazon or wherever it is they particularly go.
So that is one problem.
That bleeds in to the second which is that the traditional CIO model of command and control is probably really not going to work on them. People will not be told, they will just do it. They will agree an email and then just go and do it anyway. What do they do about that? There are various ways of looking at it. Do we try and bring it back? Using tools like BDA, do we let it go and become a service enabler? Not a topic for here but that is a big issue for them.
To wrap it up, the third issue is the fact that I do not think we are really there with applications in Cloud. The application software that has been written that is available now does not really work well the Cloud. You know that if you have gone from something, as I have in my previous working career, using even a basic software like [AT] to using a SAS application like Salesforce which is a long way behind in terms of the functionality and what we have out there in [indiscernible] software environment.
One of the things that concerns me is that with all of this technology advance we have, particularly around networking, life should really get easier for a CIO. I am conscious that it probably does not because if you are looking now at the demands of just not managing the network, but managing the performance of individual applications that might be networked globally, then that is a tough job.
Let us go to Carolyn because Riverbed is all about this. These tasks have to be done.
How much should a CIO be relying on their service provider to do these for them?
How much really has to be kept within the enterprise? Front a lot of the research we have done, and also from talking to service providers as well, most providers want to add these things as value-add to their existing services, whether they be [indiscernible] services or internet services. Are these things that CIOs are going to continue to have to drive themselves or will we get to a day when you do not have to worry about these things because they just come with your network service?
I think a lot of it is just individual choice. I doubt that all companies are going to choose the same thing because their businesses are all so unique and different. Some are going to be based upon control. I want to control my infrastructure. I want to have security behind it. Their focus is going to be on consolidating as much as possible to ultimately give them as much control as possible. So, they may choose to complete keep everything in-house and then maybe put some of their secondary or tertiary activities more into a managed service or into a public Cloud type of environment.
The challenge I think a lot of CIO are finding today is that they are losing the choice but they are being held accountable. You go and you sign up a Salesforce and you have a SAS application or HR signs up for some form of an HR application. They get the calls saying why is that not performing as if it was one of the apps that we have locally here. So they are in this precarious situation where they are being forced to be open but also being forced to be accountable.
As the CIO went from probably more recently looking at the cost saving measures around that consolidation, and how do I do this as cheaply as possible? to now rethinking it a little bit, going you know I have to factor in user experience and the performance behind it. That user experience is not only for my employees but it is also from my customers.
These are some big dynamics that are going on. With all the on-line shopping and experience we have today if you go to a website and cannot find what you are looking for instantly, you immediately turn and go somewhere else. It is very different from having a face to face interaction. That is another thing that is making people think more and more about it is not just about cost but it is the performance and the experience.
So when we look at the changes that are going on in the marketplace it is the balance of giving the IT control, the business, the financial savings and benefit along with that user experience, that are driving the CIO to think in a very different way than they have before. Those are some of the big dynamics that we see.
The last piece that I will layer on top of that is we now have all these custom application and developers that are out there that are doing their own thing. Some of the traditional technologies that we have had in the past where you have been able to accelerate a lot of the Microsoft suite or other things are all changing. That is a level of complexity that a lot of the CIOs are looking at, going how do I again ensure the performance of this when I cannot control the curve? Google Apps is a great example of that. A lot of people are moving to that but they cannot control it.
So as we look at these different things, again the standpoint from Riverbed is that whatever the choice the CIO makes, we want to give them peace of mind. That it does not matter whether they choose a SAS application or a public Cloud or a private Cloud, that we will help them accelerate all of their applications and be able to go with a centralised or decentralised environment. I do not think it is going to be one size fits all, I think it is going to be a variety based [indiscernible]
You are touching on a very good point that in the past we managed these infrastructures as separate components and if one of the applications were not performing we would call the CEO and if we were smart enough and were building a business case we would say the problem is probably with the network so we will move from 100 mig to 1 gig. If they know it is still slow then maybe we will put in 10 gig and so on. Now we are putting in 10 gig, 4 gig, and 100 gig eventually, and sometimes that does not solve the problem anymore. It is not about managing the infrastructure per se but it is about managing the application. We need to have a management system either if you want to outsource it for good reason to third party or you want to manage it in-house, you have to have a system to manage these workloads dynamically. We are not looking at this port or speed and feed on that physical switch because now that physical switch might be hosting several virtual workloads. These workload are moving from one location to a other and guess what, we want to move them from one country to another, from one continent to another.
So if you are only focussing on the physical infrastructure, having tools that are managing this infrastructure, you are really stressing out what you can do to the max.
You cannot [indiscernible] with that anymore and you need to have a more holistic point of view and say, how is this application performing, how does it behave and how can I better control it?
Just turning to Brett, the idea of live Cloud computing workloads being transferred across the world, maintaining performance, the whole thing is done transparently.
What is this going to mean to a service provider like yourselves?
I agree that one of the big issues is performance moving forwards and one of the conversations we have as CIOs is often about performance. It also leads on to scalability and flexibility. Is the platform they are designing today a platform which can support applications when in the future they may not know what those applications require. One of the big advantages that AboveNet has is that owning and operating the fibre itself, by changing the equipment on that fibre, the fibre itself is agnostic to protocols and you can get very high band widths over fibre. I also think that what the panel has said is it is not just about the cost anymore. Am I buying a solution which is fit for purpose today but will I be able to scale in three to five years’ time to the types of high band width services which we might need, with the latency we might need? If we were to go back three to five years, very few people would be able to predict the types of demands on networks that we are seeing today.
Are your customers coming to you today with demands for SLAs that do not just cover the traditional network parameters like packets lasts and latency and so on, and saying, we would like you to give us an SLA on a key business application?
We introduced what we call an agility guarantee, which is a latency SLA, less than two years ago and we are also looking at Jitter SLAs as well so it is not just about uptime anymore.
Are they coming to you and saying, we would like you to guarantee the performance of Oracle over our network?
We will give SLAs for packets, obviously the application itself. That is not something we running, maybe IBM or one of our colleagues here would be running it, but the network PCS will measure DeMark. Latency, packet loss, Jitter, all of these things are being measured and they are asking for more sophisticated tools as well to be able to monitor those systems live.
This is probably one for Daniel, but if we look at the impact of Cloud computing, the way I see certainly public Cloud computing today, that it is offered as an unmanaged service. The end user is responsible for the performance of their apps because they are defining the requirements for virtual machines and storage and all of those things.
I see the new generation emerging now where service providers, particularly the Telcos actually, want to add more value. They are able to give end to end guarantees of course because they include the network as well. However, potentially they can start to extend SLAs to performance, doing deep packet inspection on the customer’s workloads. Do you see that emerging as a requirement? Do we start to see even public Cloud computing, not necessarily private, getting that intelligent?
I think public Cloud is not just relying on SLAs, it has also not got what you would call an enterprise friendly attitude. Google would say, yes we know where you data is, it is in the Cloud. Obviously this is fine if you are just a skateboarding, funky dude from California who has got the start-up that is going to kill Face Book or whatever it is. If you are actually a proper grown-up public enterprise it is pretty useless.
I agree with your comment [indiscernible] and I do a lot of work looking at the hosting market in Europe which is obviously very focussed around Telco and Telco also have a big role in that. You see a lot of noise coming out of various telecom providers in Europe about Cloud and most of them have Cloud portfolios now, but the way I see it is it that is almost like a Trojan horse activity. They are not really that interested in the revenues, they do not really think the revenues are going to be that great. What they really are interested in doing is becoming the end point, catching the customer and being the software and service provider.
Whether that is the right way to go about selling software to enterprises, I do not know. Personally I doubt it. I think the approach you see from SIs and the folk like Riverbed who actually see a headache that enterprises have like SIFTs for example, which is a terrible problem with Microsoft, a very chatty applications, very difficult to [accelerate] with TCIP who are very successful in solving that issue. I would say that is more ways to get inside the enterprise headset rather than using public Cloud as a [indiscernible phrase].
Brett, you wanted to add something?
It was only that we had the similar situation as we are talking about, performance and the public Cloud v the private Cloud, recently with Amazon and we built a direct contact platform for some of their customers. In fact, we were the first Telco to actually enable customers to connect direct to Amazon’s account platform in the US bypassing the public internet. I think that is a trend we will continue to see so customers have options how they connect to these platforms in the future. Clearly, over a public internet you cannot give the same type SLAs, it is just the nature of the way the internet works, but over private Clouds you can.
It strikes me as odd actually that even though most of the major Telcos have moved into Cloud computing, I do not think any of them today are giving a consolidated SLA that includes both the Cloud platform and the wide area network. Of course I am not talking about SLAs over public IP because you cannot do that anyway. If you look at the big Telcos like Orange, BT and Horizon, most of their public Cloud services are being delivered over MPLS VPNs and even then you have got two separate SLAs.
Brett, is that going to change just from a service provider perspective?
Well I think contractually you will probably have two contracts in the most. What you will have is the option of using the public Cloud or the private Cloud for these platforms moving forwards. I am sure that is a trend that we will continue to see.
Actually we are seeing cautious movement towards the public Cloud. I was with a large corporation that are starting to move their non mission critical environments workloads to public Clouds. They are trying it out, they are trying the SLA, they are figuring out the infrastructure that is going to meet their requirements, but they are definitely their strategies towards a stepped approach. They are saying, let me try with some of the applications that are not that difficult and we will see how that goes.
I also think we will start to see companies some of their mission critical applications to public Cloud. In principle there is no reason why they cannot do this if the Cloud providers have the right SLAs.
Carolyn, what is your company doing to respond to the whole Cloud thing? Are there new solutions you are developing that are Cloud friendly or Cloud specific?
Yes. If you think about the normal environment where you control your infrastructure, you have the symmetric, one of our boxes in one data centre, the other the branch office. You can control that. We realised that with the adoption of Cloud technology we needed to come up with something different that would be asymmetric. We could have the box that you have inside of your office, but you needed something to be software or virtual that could be dynamically installed inside of that Cloud environment. The first thing we did is take our [indiscernible] optimisation steelhead technology and made a Cloud steelhead version. So back to your point of companies that want to operate out of the public Cloud, they can now put in the Cloud steelhead and get the exact same performance as if they owned the equipment themselves. That was a big step for many of our customers to be able to have that option because it did not exist before.
The other thing that we have done with things like Amazon and others is, for some of the less mission critical, back-up is a great example. I do not think you will find a customer on the planet that really enjoys doing tape backup, but they have stuck with it because despite a lot of the other alternatives it is still something that they have kept in their process and the way that they run their business.
The other thing that we did was took a technology which we call white water and you can backup to the Cloud.
The other thing we tried to do for companies that do not want to change their backup software regardless of whether they are running Symantec or Tivoli or something else, they do not have to change this. We run underneath that so that if you are running that software, that is great. We will just go in behind it and then you will be able to backup directly to the public Cloud.
These have been hugely enabling technologies for us.
The other piece is just around the management. A lot of times into the Cloud they cannot do load balancing or other traffic management so we also have another technology that gives companies the visibility to understand, a) what is happening performance wise and, b) to be able to go in and make corrective changes, even if it is a public Cloud environment.
So to us, these are big enabling technologies that will power through the barriers that have been stopping companies today.
I have had a signal that I think we are in our last five minutes. I would like to open it up to the delegates here. Any questions for our panel?
While you are perhaps thinking of a question, let us move on to the last topic. This is one that I touched on when I was doing the opening piece here. In terms of helping CIOs in doing this role of becoming a champion for broader innovation within the business, is this an important role for your companies today or do you leave them to go off and do this on Google themselves?
You can do it on Google of course, but one thing you might find on Google is the IBM Institute Business Value report. This is something that IBM publish once a year that sets down leader of the world, politician, technologists, etc. and looks at the different trends and tries to spot where are the pain points going to be in the next few years. In which sector? This is why IBM came up with a smarter planet concept where we would be dropping integrated systems to optimise cities and how we manage traffic, or we are looking at healthcare and how to develop a better, smarter healthcare system. So, the IBM Institute Business Value report is available of course on Google and others. Some of the reports and trends we have to be looking at are transcending the technology itself and they can be related to society or to different perspective like that.
One of them that we are looking at very actively right now is Africa. We are seeing a lot of technology that we are developing now is being adopted very early on in Africa.
There is a lot of youth, a lot of good energy over there so some of the trends and technology that IBM is putting forward is built towards that objective.
Carolyn and Brett, do you see this as a role of your companies as well?
We do. Just in the most simplest form, CIOs are now being tasked to think more about how they can help their companies sell more products and increase their revenues. So for us, whether it is working on a website and being able to conduct more transactions and sell more, or for them to be able to dynamically set up infrastructure maybe during holiday buying seasons or for events such as the Olympic, to be able to have them set up and take things down in a very cost effect way that helps their company grow their revenue and wallet share with people.
For us, we have enabled a lot of the CIOs to go and have that conversation about not only do they have a better infrastructure, but they can help the company sell more and that has made a big difference in how they have positioned themselves in the organisation.
Bret, your light is on so you have something?
I was just going to add to one of the points that Daniel made at the beginning about the fall of IP transit. I think one of the elephants in the room is what if the price of IP transit continues to drop? Yet one of the other conversations we were having today was really about performance flexibility of your network and whether you can get that type of performance over IP transit. I think it is very hard to sustain both if enterprise customers are looking for the lowest cost IP transit provider, they may be using a tier two provider which is reliant on getting its routes through public exchanges. This is not the way that you are going to get the performance you need to enterprise clients’ performance to Cloud based services. I think that there has to be an acknowledgement that IP transit is not a commodity if you are going to be using it for mission critical type applications.
Actually there is some stabilisation in pricing. If you look at IP transit above 10 gig, prices drop as volume increases until you get very large amounts of band width and at very large amounts of band width, there are only certain tier one providers that can provide that. So there is an unusual phenomenon where you get to the very large volumes of traffic and suddenly things like your routing tables become much more important, it no longer becomes a commodity. I think it will also become very important to customers who are putting more of their infrastructure on Cloud based services. We are seeing the phenomenon where customers who used to ask the price now want to look at your routing tables and they want to look at the latency of your network between the key locations, data centres and services.
One very quick question if anyone has one.
Okay, I would certainly like to thank the panel and hope that the discussion has given you some ideas.