Network Optimization Made EasyBy Channel Insider Staff | Posted 2007-09-12 Email Print
Duane Wells, CEO of solution provider ITI Solutions, joined Ziff Davis Enterprise Editorial Director Mike Vizard for a recent Changing Channels podcast. A transcript of that interview follows.
Mike Vizard:I guess every time you go into an IT organization, it always seems like the fingers are pointing hard and heavy down at the network department because it's always, you know, the applications are running slow, the IO for the servers and the storage is slow because of some mysterious thing that must be the networking guys' fault. A) How prevalent is that today and, B) Is that going to get more prevalent as we go forward as far as you see with yourf interaction with your customers?
Duane Wells: What we've found going into customers is that it's not always just an issue for the management of these companies to just figure out whose responsibility it is, but to also try to find a way to help them play nicer. And we've found that, in terms of right now, the IT departments fighting with each other, that it really helps to have something to help them play nicer together. And that's more of the issue that we've seen now more than just finger pointing. I think finger pointing is a great thing, I mean in terms of identifying if there's a problem. But in terms of what the real issue is, I think that today and especially as density — in terms of the applications that are housed by these companies — increases, it's really going to be a matter of trying to get them to play nicer together in resolving issues rather than just it being finger pointing. And we see that as a trend.
Vizard: And then driving that density is things like virtualization and things like data center consolidation.
Vizard: It basically feels like we're trying to increase the utilization rates of servers, but when we do that, we put more pressure on the network.
Vizard: OK. What kind of opportunity does that create for you guys? How are you trying to, you know, actually make some dollars out of that?
Wells: Well, you know, I haven't done enough research to know exactly what's happening across the country and around the world. But I know that in the region where we're marketing in terms of our prospect base, we are the only company that has decided to make network analytics a focus. As I said a few years ago when we were looking for opportunities in the industry, I really started to feel like with the proliferation of all the different IP-based applications, host applications, virtualization and VOIP — I looked at the need for a company to actually understand better in terms of the way the traffic flows on the network. I looked at an incredible opportunity. So what we're doing right now is we are really, really focused on trying to become a solution provider that can help companies understand what hardware and software solutions they can utilize to get better visibility on the network. And that's really been our core focus.
Vizard: And what are you using to actually drive that? What vendor have you hooked up with?
Wells: The primary vendor that we've been working with so far has been Network Physics. We've been very happy with them. Their product seems to be, across the board, the best bang for the buck, and it gives the most advanced level of technology in terms of analysis. And they're great to work with. So we've been pushing them primarily. You know, my vision for the company really has been around trying to bring the best solution for the customer. So we have kept our eye on different companies in the industry. We talk to them, we've done apples-to-apples comparisons in … we're constantly setting up labs here at our company to try to do these sorts of competitive analyses on our own so we're ready. But primarily, it's been Network Physics.
Vizard: Now, a lot of times you talk to solution providers and they say they have a hard time getting paid to do things like analytics because it's almost like the customer expects them to do that just to bid for the account. So how do you make sure that the customer sees value in this proposition and is willing to pay for that service versus treating it as, you know, almost like a chip to ante up to the game in the first place?
Wells: Well, that's a great question, and I think I can answer it in two ways. One has to do with how we sell it, and the other has to do with the maturity of the industry. I think that right now, it being a very young industry in its own right, it is difficult to get companies to ante up and to justify the expense. Now, generally, we've found that it's a slam-dunk or it's, you know, it's a long crawl. There is not a whole lot of room in between. Because when companies are feeling the pain and you come in and you actually address the problem in a way that they can relate to and understand and it's really going to help their business and their bottom line, it's really a no-brainer. So everything that we do is proof of concept. We're always showing the customer what it is that they can do with the technology that we're trying to sell prior to them opening up their wallets and paying us anything at all. So now, that might be how it is today. However, I think that in terms of how the industry is maturing, and as it starts to take hold and companies start to realize that this is, in its own right, its own sort of technology base that they can start to take seriously, that it'll become easier to get companies to start to pay for these kinds of services.
Vizard: Do you think this is going to become a bigger problem? Because I look at the next generation of applications and I think about, you know, people are trying to take business processes, push them over the Web. They're trying to use things like Web services and XML to do that. And all that creates more sensitivity to the latency of the applications. It sounds to me like, you know, density is the first wave of the problem, but latency is right behind it.
Wells: Sure. Well, you know, in terms of the maturity of the industry itself, I think that it's not just about the density or the latency. I think that as the application density itself proliferates, latency will be an issue. And you know, our approach is really to address the conversation that the application and the server infrastructure and the network infrastructure are really communicating with the end user. So I think that it will make a difference, in terms of latency, as the density increases. I definitely agree with that.
Vizard: Do you partner with other solution providers out there who are specialists in the networking gear, or do you sell that yourself? Because part of the challenge here is trying to figure out the relationship between the application teams and the network and server teams and, you know, that can be as many as three different solution providers.
Wells: Yeah, we do … as we are finding that companies want us to be partnered with those companies, and as we find that, we may go into a company to do an engagement and they may have a certain switching fabric that isn't necessarily giving them the best bang for their buck. Or it's not allowing them to accomplish some of the needs that they have for their infrastructure. I think that we definitely like to go out and partner with companies to do that. When you take the larger companies, it creates an interesting sort of a situation because a lot of the vendors themselves have their own ideas about how to deal with these problems that may be more traditional and different from our approach. So it does become a challenge sometimes partnering with those other companies. But we do in fact, as it comes up and becomes necessary, we hook up with these companies to go in and attack the engagements.
Vizard: Now, don't they all kind of have some kind of agenda of their own right? Because, if I'm the network hardware guy and I'm selling my solution, the problem is going to be, you know, "Here's this lovely LAN accelerator or application accelerator that you should buy from me." Of, if I'm on the server side, you know, my answer is, "Buy my faster Ethernet card or my faster IO solution for storage." How often, when you go into a customer and you see these problems, is it really a hardware problem versus a configuration problem or some kind of pilot error problem?
Wells: Well, we haven't found that there is a particular trend in terms of the kinds of problems that we're facing, but that is a good question: How do you handle vendors that don't necessarily want to work with you? And you know, generally speaking, it's just been a process for us. We've had to work with them and show them that we really have a holistic approach and that we really are trying to work with them to help their business with the customer that we're working with. I have my own theory of sort of how that might play out. I think that it will become an increasingly difficult challenge as things move forward. But up until now, it's worked out pretty well for us.
Vizard: Do you ever see people thinking about actually trying to manage this at a more holistic level? Because one of the things you see out there is, the IT organization is wrapped around, "Here's my storage discipline, here's my server discipline, here's my network discipline." And as a result, services companies then have three different practices. But you know, as we increase and we go forward, it's a holistic conversation, and anytime you touch one element it has an adverse impact on the other elements. So are we going to see maybe a more unified approach to network systems management or data center management as we go forward?
Wells: Interestingly enough, the question that you present actually has influenced our model. One of the reasons that we really, really like the Network Physics product so much — and there are a few other products out there that do this — is that we are completely agnostic. The model that we have, because of the question that you presented, we have decided that we want to try to stay as agnostic as possible. So no matter what, we can approach each engagement holistically and we're not really dependent on any particular technology. Because I think that is something that's going to become more of an issue, especially as you start to spread out and diversify and practices that have traditionally been under one umbrella become unique and on their own right. So we're going into it with the idea that it's going to be a, it's a holistic approach and that we are going to be completely agnostic to whatever particular ways the customer may have their infrastructure put together so we can really look at the goals of what they're trying to accomplish and then really look at end-to-end analytics and application discovery.
Vizard: And do you have any secrets for helping to ensure profitability in that model? Because, again, the first thing a fellow solution provider would say is, "Well, you know, essentially that means I wind up bundling services together and then my profitability drops down." So how do I get down to that holistic approach and still do that in a way that maintains profitability?
Wells: Now, do you mean in terms of the … you mean sort of the customer itself?
Vizard: No. I mean your profitability as a company.
Wells: Well, you know, I'm a firm believer that if we approach it the right way, the models in terms of profitability we've looked at have looked very good as far as our approach goes. And we really believe that if we stick with that approach, there are going to be some times where it's going to be a little bit thinner for us. But I think in the long run, in terms of us trying to make a name for ourselves and the approach that we're taking being really, really pure in terms of the fact that we're looking at what the customer's going to need in order to take best advantage of our technology. So I think that it will be a struggle for us from a profitability standpoint sometimes to adhere to that model. But in the long run, I really think it's going to pan out.
Vizard: Do you perceive that at some point this analytics tool becomes the foundation for a different type of managed service that you might provide your clients, which, in turn, creates more of a profitability model for you because you're not treating each client as a one-off engagement, per se?
Wells: When you look at the Network Physics product in specific, one of the things that is so nice about it and one of the reasons we like it so much is that you do have multiple approaches to multiple problems in terms of how you go about doing your optimization and monitoring. The Network Physics tool itself is really, you know, we really consider it the Swiss Army knife of network optimization tools. And when you're going in and you're trying to find out what a problem might be, you may go in and look at conversations, but you also might look at how your traffic is behaving. So we go into an engagement and, in specific, we talk about profitability on the managed services side. We actually will go in with the Network Physics tool and you may have a situation, let's say, where you may be monitoring a certain level of traffic and you're monitoring it for up and you're monitoring it for down. And you might be monitoring it for latency. But what you're not doing is you're not monitoring it based on characteristics that are specific to that type of traffic flow. So you may have a remote office that does all their sales reports on the 15th, and they're going to transmit a large amount of bandwidth. And you're expecting that to occur, and you want to know that that happened. With the Network Physics tool, you can actually go in and say, "I want to monitor this circuit at this time for this sort behavior. If that behavior doesn't occur, I want you to let me know." And by bringing that type of analysis and monitoring to our customers, a lot of times they find that very, very useful. And when it has to do with their bottom line, they're willing to pay to have those kinds of alerts put together and those kinds of reports put together. So we feel really good about using the tool to sort of create specific revenue models. And it's really very young. So we're just sort of seeing the beginnings of what it could mean. But we really think it's going to be great in terms of creating profitability for us going in and allowing us to really customize the way that we optimize and the types of alerts and reports that are generated for customers that they find a lot of value in.