Open-source networking doesn't require a guru

Vyatta CEO and chief strategy officer on the open source router company

Open source router company Vyatta debuted earlier this year with a Red Hat-style alternative to Cisco and Juniper offerings: the Open Flexible Router, an open source-based WAN router and firewall stack, freely downloadable, with service and support offerings available for purchase. Since then the company has generated buzz in the network industry, while releasing products such as a pre-installed appliance-like version on Dell servers. Vyatta CEO Kelly Herrell and chief strategy officer Dave Roberts recently told Phil Hochmuth what Vyatta is, and is not, and what it hopes to become. (The following is an edited transcript.)

How much future is there in being an open-source networking company, given that a lot of what you're doing is packaging free technology that has been out there for some time? Is this something people will do themselves?

Herrell: Companies buy solutions. I don't mean to sound trite like that, but when people look for solutions, they're looking for something that has the best price/performance for the job at hand. We are a solutions provider. I don't know of too many CIOs who would look fondly upon their network teams if they were sitting in labs trying to compile and debug code, or something that was a standard function in their network. So as a solutions provider, we can give buyers what they want: continuity of the product, maintenance, a road map. You're not going to get much of a road map out of an open source project. And all the technical support and service that they rely on to run a network. So we're an open systems alternative to a proprietary approach. So from that perspective, we are a solutions alternative, and we believe that has just as much longevity as a proprietary solution.

What are the advantages of open source from a competitive standpoint?

Herrell: We pull from various parts of the open source world, and contribute back, of course. Everything from Linux, and XORP [eXtensiple Open Routing Platform -- the open source routing software on which OFR is based]. The advantage here is we're standing on the shoulders of giants. Many of the components have been weather tested in many other environments. So we're not coming in at the fundamental ground level where it's a systems theory. So we hit the ground running. And then from our perspective, our job as a solutions provider is to continue to make quick incremental improvements to the solution just to continually advance the state of the solution that we offer.

How come XORP, and open source routing in general, has not taken off as widely as open source computing or application platforms, such as Linux or Apache?

Roberts: We're fond of saying that networking started out open source. The first networking stacks were open source stacks with either BSD, or people would run [networking] software on Sun. That was a common way to get yourself on the Internet in the 1980s. Then the networking market swung closed source, and went through a period of extreme growth through the 1990s. I think now it's reached a level of maturity where people are willing to go look back at those open source solutions.

The market has to be accepting of what you're doing. We're at a point now where the market has come to the conclusion that open source is good. Customers want to see more open source alternatives in a variety of product categories, not just computing.

To be a Vvyatta customer, do you have to be a hacker or open- source guru, or have one on staff?

Herrell: Let's say you pick up the Vyatta appliance, or just the OFR software. When you push the on-button of the machine, it boots. When it finishes booting, it is a router, with a CLI and a GUI interface. So the comparisons to Red Hat for us are very apt, in many ways -- the subscription model, and the leverage of open source. Where they differ, is that what Red Hat provides is an operating system, then you have to load apps on it, and do all that kind of thing. From Vyatta, the product you get is the same, from a user standpoint, as a traditional router or firewall. You plug in the cables, you hit the on-button, you configure it, and you're done.

Roberts: You absolutely do not have to be a hacker to use this product. This is really designed for your average network manager who is comfortable with a Cisco or Juniper product today. They can fire up our products and find themselves very comfortable. One of the things about why open source networking had not, until Vyatta, really caught was because to a certain extent, the solutions and open source stacks that were out there -- XORP and others -- do rely on users to be a little bit more of a hacker to deploy them. You still have to download them, you have to run them on your Linux distro. You still have to understand Linux, because it's not like you get a full environment. When you boot up your raw XORP-based system, you have a set of processes running on top of Linux. You have to know Linux commands to maintain the system. That's where a lot of our value-add is; not just taking XORP and plopping it on top of Linux.

Page Break

What are customers' biggest reservations about going with an open source networking product?

Herrell: Change. And I would qualify that by saying that in any market there are different types of adopters. For those who are resistant, that's fine. We'll evangelize and proselytize, but we won't try and force someone to do something they don't want to do. Our job has less to do with dealing with objection and more to do with understanding where the pockets of adoption are.

Roberts: I think everyone gets what we're doing. We haven't talked to anybody that doesn't get it or doesn't see some benefit in it. I've had major Fortune 500 corporations saying, this is really interesting. They've also followed that by saying, I'm not sure I'm ready for it. I'm not sure our organization is ready for it, but I understand it and I understand the benefit of it.

What do you tell CIOs when you talk about Vyatta's road map?

Herrell: We tell them we've got our 1.0 release out there and that 1.1 is around the corner. They should expect it to look like any commercial product. And from that perspective, the road map includes feature advancements and performance enhancements. What is new is the way we come up with the definition of what needs to be in those incremental advancements. That's where we get to leverage the community. We get to leverage their insight and their requests. We don't build something because we think it's a neat-o idea. We build something because the market is telling us they want that.

So who is ready for open source routing?

Herrell: The first adopters are generally categories, are organizations with nimble budgets and nimble deployment models. Who fits under that? Well, SMBs, service providers. People who aren't going to require a long, protracted formalized product review, but rather, a customer that will say, hey, I have a need. You have a solution, I'll plug it in. If it doesn't work I'll take it out.

Roberts: These are typically organizations where there is some empowerment by technical people to make decisions. As opposed to large central planning committees for technology buying.

Herrell: Yes, Stalinist regimes need not apply. But back to what Dave said, I haven't heard any senior IT manager or CIO say, no-way, no-how. What they say is, it's interesting, I'm going to watch this. The bottom line is, no wonder it's interesting, it has two basic advantages -- economics and control. How do you say that's not interesting? If you say, I have a very significant network budget, and you can stretch that farther. Or, if it gives customers more control over what I deploy, how I deploy it, and when I deploy it. Those are good things. Customer A might say I'm taking a wait-and-see attitude, but I don't' disagree with the approach.

For organizations that are in wait-and-see mode, what do you think they're waiting to see?

Herrell:I think the funny thing is that what they need to see already exists; we're just in the process of communicating to them that it does exist. What people like to see is that somebody else has done it. It turns out that a reasonably good-sized number of organizations have already done this, and now have production networks running Vyatta. And it's up to us to explain that and show the proof, if that is the pressing item for them. For the most part, I think that's it. We don't get resistance to the idea; they just want to make sure they're in good company.