Egenera CEO stresses openness in competing with ‘Cisco, Cisco, Cisco’

Egenera CEO Pete Manca says the company was 10 years ahead of its time in targeting the converged data center infrastructure, but its smooth transition from a hardware-focused company to a software-oriented one in recent years has enabled it to stay in the thick of what’s now a booming market pursued aggressively by Cisco, HP and others. As part of our ongoing IDG Enterprise CEO Interview Series, IDGE Chief Content Officer John Gallant spoke recently with Manca about Egenera’s strategic shift, its partners and competitors, and a big new product announcement. 

PRODUCT ANNOUNCEMENT: Egenera jumps into cloud management

John Gallant: What’s Egenera’s mission today?

Pete Manca: The mission from day one, and now we’re 12 years old, has been to simplify the data center, and we do that through converged infrastructure. 12 years ago when we were pushing converged infrastructure, everybody kind of scratched their head and said, “What do you mean? What is it?” Now with products like Cisco’s UCS [Read Network World‘s review of UCS], IBM’s PureFlex and HP’s Matrix there are a lot of examples in the market. But it’s really the technology that we invented 12 years ago, and very simply, what it is is the combining of resources connected to a common networking fabric. In our case it’s Ethernet, which is really the case for most places now, most vendors. And it’s the pooling of storage, networking and server resources so they can be dynamically assembled at the time when you need them. You do that by inventorying your resources with a little magic of virtualization, and not necessarily server virtualization like VMware, but I/O virtualization, really virtualizing the fabric so you can piece together those resources when you need them. It reduces your costs significantly from a hardware point of view, you need less resources, less port counts. It reduces your admin costs significantly because you need fewer people to manage these machines. And it gives you a much more dynamic environment. So our vision, really, is delivering converged infrastructure to the market, helping the customer simplify their data center. The real differentiator for Egenera is we are really the only open converged infrastructure player in the market.

BIG SHOTS: Catch up on all of our IDG Enterprise CEO Interview Series articles

Gallant: Open meaning what?

Manca: We run on everybody’s hardware: IBM, HP, Dell, NEC and Fujitsu. We apply our converged infrastructure management software to those platforms and we can turn those into a converged environment.

Gallant: Let’s take a step back, because I think people who know Egenera from the beginning know it was essentially a hardware company. How has that mission evolved, and did you drive that change?

Manca: One of the things that I did, my main goal when I came in here, was to transition us from hardware to a software company. We had been a hardware company for the first nine years of our lives and were very successful. We had several hundred customers, we ramped the business very quickly to $100 million, but we could never attain profitability. We couldn’t get the business model correct. As other vendors got into the market and blades became more of a commodity, it became even more difficult for Egenera as a hardware company to get the business model to work correctly. So we had to change something dramatically. That was my job when I took over as CEO two and a half years ago. What we did is we took the core IP of the company, which was PAN Manager software, which ran on our proprietary systems, and we poured it into open systems. We started with Dell and added Fujitsu, then we added HP, NEC, then IBM. So now we’ve been able to take that, really the crown jewels of the company, and provide that capability on open platforms. Since then I’m very happy to be able to say that we’ve been profitable for the last two years. We’ve got the business model corrected. We’re generating cash, profit, and now we’re looking to really extend the functionality.

Gallant: How were you successful in getting all of these companies to embrace this software? Don’t they have competitive offerings of their own?

Manca: They do. I think there were a couple main drivers. The first was Cisco. Cisco coming in with Cisco UCS [Unified Computing System] and the VCE initiative really became a major competitive threat to these companies. Cisco has an enterprise sales force, they understand very well how to sell this product and they did something that the other vendors didn’t do, only Egenera did, which is they started from scratch. It’s very difficult to create a converged infrastructure product when you’re piecing together components that are either homegrown or from acquisitions, and then you try to put that together under an umbrella and call it converged infrastructure. The whole benefit of converged infrastructure is simplicity. In some cases, some of these vendors are creating a more complex environment by piecing together components that were really never meant to work together. Cisco, like Egenera, developed from scratch and did a nice job doing it. Some would say they copied our model. We were certainly a very good solution for them to do that. We have probably the most advanced high availability and disaster recovery capabilities on the market and some of the other vendors don’t offer those capabilities.

Gallant: I want to go in some depth about how PAN works, but let’s stay at the industry level for a second. Who do you compete with on this?

Manca: Cisco, Cisco, Cisco. Cisco number one, Cisco number two, Cisco number three.

Gallant: What is it about UCS that you compete with though? What does Cisco bring at the software level that competes with what you’re offering in PAN?

Manca: It goes back to the simplicity and starting from scratch. They bring a set of functionality that works well together and it’s integrated from a single vendor, which is a good strategy; obviously, it’s one that we chose. And they’re everywhere. They have a very good enterprise sales force. That maybe is one of the main differentiators. If you look at some of the other vendors, some of them grew up as box vendors.

CISCO SUBNET: An independent Cisco community

I don’t want to sound derogatory in any way, but they’re used to selling boxes, not solutions, not software. Cisco had a different mindset. They understood how to go into the data center and sell a high-end solution. That put a lot of pressure on some of these vendors.

Gallant: So in terms of competitive differentiation versus Cisco, what do you help these partners bring to the table that they don’t have right now?

Manca: Certainly openness. That doesn’t necessarily benefit each individual partner, but it benefits the channel, it benefits the end user customer because they now have a choice. Cisco certainly is almost back to the future for us. They look like what we did 10-12 years ago, proprietary stack of hardware and software. So that’s a real differentiator. The other is we have very enhanced high-availability disaster recovery capabilities, where we can automatically failover any resource that fails locally and remotely. That’s patented technology. It would be very difficult for somebody to replicate that since we do own the patents on that. When we get into talking about some of the future stuff, we’ll talk about some other areas that we’re going to be differentiated in as well.

Gallant: In terms of the types of customers that these solutions are geared toward, is there a different set for Cisco and a different set for Egenera?

Manca: I don’t think so. These types of solutions work very well for service providers. They work very well for mid-enterprise, SME, all the way up to just below the large enterprise, because those are the folks that struggle the most with resources. If you go to a large Wall Street bank and they have 1,500 engineers working on IT, a lot of times they’ll try to cobble this kind of stuff together themselves. If you go down a tier, the $1 billion-$5 billion range, and these folks maybe have a handful of people in their IT department, they don’t have time to be cobbling things together. They want a solution that you can wheel in, turn it on and be up and running within hours, and that’s what PAN Manager provides them. So we find a lot of success in that market. Service providers I mentioned, and we’re seeing a lot in health care.

Gallant: So are there other competitors? You’re talking about I/O virtualization and I know of companies like Fusion-io. Is that too small a piece of the pie?

Manca: Yeah, a different market. It’s solving a different problem. There really are no other competitors for this other than Cisco, and of course some of the vendors’ homegrown technologies that we bump into every once in a while.

Gallant: Are initiatives like OpenStack or CloudStack trying to achieve the same goals?

Manca: They’re higher layers in the stack. We sit below that. In fact, an OpenStack, CloudStack might sit right on top of our product and get the benefit from our I/O virtualization and our management capabilities underneath the cloud management solution.

Gallant: I want to talk a little bit more about partners before we shift to the technology. You mentioned a lot of partners and they’re all the big names in the industry. Who are the key partners though?

Manca: Right now it’s HP and Fujitsu, but I expect that to shift over time. IBM just came online last week. NEC [general availability] is next month, so they really haven’t hit the market yet. We have global partners like HP, IBM, Dell, and then we have more regionalized partners like Fujitsu who is very strong in EMEA, through Fujitsu Siemens, the previous incarnation. And we have NEC who is actually the No. 1 Intel server vendor in Japan.

Gallant: Are there people out there that you’re still trying to bring into the partner fold?

Manca: We are talking to one or two more, but I’d say they’re at very early stages. I would think of them more as regional players.

Gallant: Do you use them as your channel partners or do you also have a channel of your own?

Manca: Fujitsu and NEC resell our product regionally. NEC has a global agreement but they’re really a regional player, they’re really Japan. For the most part we’ve built out a channel program ourselves. We’re not relying on these vendors to be our channel. About a year and a half ago, we started a channel program and we’re just starting to see the fruits of it. We have a substantial list of channel partners in very large companies like SHI. We’ve got regional channels as well, and we’re starting to see leads come in.

Gallant: So can you give us some perspective of growth? I mean, how is this strategy doing, how is it changing the company’s direction?

Manca: Software growth has been very, very rapid. We just closed 2011 and grew 407% year over year in the software business. Q1 just completed; it was a record quarter for us and we grew 100% compared to Q1 of 2011. So the growth continues even as the numbers are starting to get bigger now. You could argue before maybe it was the power of small numbers. Now the numbers are getting bigger and yet the growth is still there.

Gallant: How does that growth fall? What percentage is service provider vs. enterprise?

Manca: We’re about 35% service provider. The rest is split between healthcare, finance and government. Government may be No. 2 at around 30%.

Gallant: Are you seeing different growth rates in the service provider vs. enterprise segments?

Manca: Service provider is growing faster. What we’re seeing though, is a tremendous push for private cloud more so than public cloud. And that’s because of the nature of our solution. We don’t typically go sell to Amazon or the large public clouds, that’s not our focus. So maybe that’s a self-fulfilling prophecy, but we are seeing a big push for private clouds, folks that want the benefit of the cloud but want to own the resources in-house or host them at an ISP, but still have control over their own resources, not being shared.

Gallant: OK. So this is primarily a private cloud kind of solution?

Manca: It is. It can be hosted or be inside the data center, inside the enterprise.

Gallant: So Pete, go into more depth about PAN. How does it work?

Manca: PAN Manager has three major buckets of capability. The first is discovery and provisioning and what that is, is the ability to go out and discover other resources that are on the fabric, dynamically provision them — and this is where some of the I/O virtualization comes into play. Do multi-tenancy, so you can share those resources across different end users, layer that with security and charge back. A typical management capability that allows the customer to provision and build out a data center. Then from there you can add high availability, which gives you the ability for any resource that fails to fail over dynamically, whether it be a server, a link to a disk, a network link, it doesn’t matter. Then the third bucket is the disaster recovery, which is taking high availability and extending it to a remote site. We can take an entire PAN environment and within a matter of minutes replicate that in a remote site across the world and have the customer up and running.

Gallant: A lot of our readers are talking about private cloud and perhaps under the perception that they’re building private cloud. What is it this enables you to do that most people would struggle to do on their own if they’re using virtualization and they have a set of security tools and management tools today?

Manca: If you think about the way clouds are provisioned today, they typically run on top of a virtualization engine, be it VMware, Citrix XenServer, etc. It’s a management cloud/management tool sitting on top of virtualization that allows you to provision virtual machines. Well, the whole world isn’t virtual. In fact, when you’re in the enterprise, a lot of the applications are still running physical. And so one of the things PAN Manager allows you to do is provision a physical cloud. We can do the same capabilities with physical servers. We don’t need the virtualization. We don’t need a Hyper-V or a VMware. If the customer wants to choose that, that’s fine, we can manage that environment as well. And what we can do, and what we often do in a lot of customers that they find unique, is we can allow them to mix physical and virtual inside the same cloud and have a single application. I’ll give you an example. Brown Shoe is a customer of ours and they run SAP on PAN Manager. SAP does not like to run all virtual. The database, the back end is physical, and the front end stuff may be virtualized. Well, the customer could have two different systems and two management interfaces to manage that or they can choose PAN Manager, and they can manage both a physical and a virtual cloud from a single interface.

Gallant: So is this competitive with VMware?

Manca: The messaging gets very confused because if you read VMware product marketing material, you hear about HA and DR and provisioning. But they’re focused 100% virtual. We’re focused on physical or physical and virtual. Probably 60% or 70% of our customers run some kind of virtualization engine on top of PAN Manager. If you think about the benefits, where the virtualization engines stop, they can failover virtual machines, they can do disaster recovery in virtual machines, but if a physical resource fails that’s where their management reach ends. In VMware’s case, or with any of the hypervisors, if you have two machines and one fails, they’ll fail the virtual machines over to the one machine that’s still up and running, but that machine now has taken on more risk. It’s taken on more capacity. PAN Manager, underneath the covers, will bring a new server online, make it look like the one that failed, it’ll come back up and almost look like a reboot, and all those virtual machines will dynamically flow back to other system, then balance [all your load again].

Gallant: And could you do this in an environment with multiple different hypervisors?

Manca: Yes.

Gallant: And it still works across all of those?

Manca: Yes.

Gallant: Which is something those others would struggle to do?

Manca: That’s exactly right.

Gallant: What about in a hybrid environment? Does this work if you are running a private cloud but you’re also moving workloads up into the public cloud?

Manca: Do you want to start talking futures now?

Gallant: Please, whenever you’re ready.

Manca: The most significant feature we’re going to be talking about is a cloud product [read more about the new products here]. And what’s going to be unique about this is it’s going to allow the customer to provision physical and virtual clouds from the same interface. It’s going to be the only cloud product on the market that can do that. So it will be a standard cloud provisioning engine, a self-service portal with the ability to charge back to calculate for resources, but now the customer through a drag-and-drop palette can create applications that span physical and virtual servers and do all that from a single interface.

Gallant: Does that require partnerships with service providers, or it’s any service provider you would want to work with?

Manca: Any service provider you’d want to work with. It’s going to be a very interesting technology stack. The other big pieces of that release, we’re scaling now basically to unlimited servers.

Gallant: I was just going to ask you about scale.

Manca: We’ve been limited to up to 64 servers and that’s been a factor of what the server vendors themselves support in a domain. Like HP supports four chasses in a domain. We broke through that. And so we’re going to qualify up to 256 servers, but it’s theoretically unlimited. We can qualify larger if we have larger environments, but we’ve broken through those physical barriers that the vendors have. We’re also adding a storage management product. And we are adding a plug-in to the VMware Orchestrator. So again, if the customer has decided they want vCloud and they want VMware as their management environment, we can plug into VMware Orchestrator and allow that customer to build workflows across physical and virtual machines on the same interface. When I took over, I had a three-step plan. Step one was: Get on as many major platforms as we can. That step is now complete. We may add one or two more here or there.

Gallant: You’ve got a lot of [that front] covered.

Manca: We’ve got the major ones covered. Step two is all about scale and virtual and physical cloud management. And step two really is more of a product cloud play in the ability to manage physical and virtual from the same interface, any way that the customer wants it. We want it to be open, whether it’s our cloud director or if it’s VMware or other cloud tools we may choose to integrate with later on. Step three is of a hybrid cloud, bridging to the public clouds from the private cloud. There’s a lot of buzz out there about that now, but I think there’s a lot of technology that has yet to be delivered that makes that doable. So I think we’ll be right in the mix for that.

Gallant: Does that change the competitive landscape?

Manca: Yes and no. Everybody’s rushing toward similar types of capabilities. If anything, it makes it more intense with some other vendors out there that we may find ourselves competing with, some of the more traditional software houses. You look at even EMC with their [VSPEX] announcement a few weeks ago? They’ve already got Vblock and VCE, now they’re going to come out with their own reference architecture for this type of solution, because it makes sense. The vision we had, the idea [founder] Vern [Brownell] had 12 years ago, was really spot on. It was 10 years too early, but it was really spot on. The market is validating that now.

Gallant: You mentioned that one of the upcoming announcements was a storage management piece. Can you talk a little bit about what that will be?

Manca: Sure. In the original BladeFrame product we managed storage, we virtualized the storage interfaces. When we went to the open model, we changed it from management of the SAN to management of HBAs, because that’s what we were given by the server vendors. A lot of customers said, “I want you to go back to the way BladeFrame did it,” which is, “I don’t want to think about HBAs, I just want to pool the storage LUNs that I can allocate dynamically.” So we’re going back to that model. We’re going to give the customer the choice that if they want to manage HBAs like they do traditionally, fine. If they want to manage pools of LUNs, they can do that as well.

Gallant: So give our readers a sense of the cost and the implementation process. What do you have to do if you’ve already got a highly virtualized environment?

Manca: You can be up and running in a matter of two or three hours. It’s very, very simple. One of the other changes we made as we went through the platform is we became completely out of band. With our BladeFrame product we were always in the data path and had device drivers you had to get. Red Hat device drivers, Windows device drivers, Solaris at the time. Now we’re completely out of band, we have no code in the data path, no drivers you have to install. As far as cost, we sell it in three different skews: We sell the base product; the discovery and I/O virtualization and the provisioning engine as the base; and then we upsell high availability, we upsell disaster recovery on top of that. Cost typically runs in the few thousand dollars per blade kind of number. It really depends on the mix, how many HA blades they want, how many DR blades they want. With service providers, even that pricing model can be custom and it could be a pay-as-you-go kind of model as well.

Gallant: So put some other numbers around this. When somebody moves to this typically, what kinds of savings can be expected, how do you measure the impact of this?

Manca: It’s typically a TCO analysis, and we see savings around 50% in terms of hard and soft dollars. We have a pretty sophisticated TCO calculator and your results may vary based on how good your internal practices are. You may or may not get a lot of soft dollar savings, but you’re going to get some either way. You get hard dollar savings because you need less — because we virtualize the I/O you need fewer NICs, you need fewer HBAs, you need fewer switch ports, so you just need less infrastructure. You get the soft dollar savings because it essentially manages itself. You set it up and you go, and you don’t have to be around if there’s a failover, PAN Manager just deals with it and the next morning you come into work and you find you may have to physically replace a blade, but you didn’t have to do anything that night.

Gallant: So who typically makes the purchase decision about PAN Manager?

Manca: It’s typically a senior-level IT person. Because it’s a transformational-type technology it tends to go all the way up to a VP level. One of the places where we get some tremendous savings, that doesn’t jump out at you from looking at the technology, is on software licenses. Because we’re reducing the number of servers you need, you need less software licenses, but also because we fail those licenses over when a piece of hardware fails, you don’t need to have double licensing anymore for a high availability disaster recovery. We can push those same licenses over.

Gallant: We hear a lot about private cloud, and I think some of the things that people don’t understand about private cloud is the scale issue. It’s not just virtualizing, it’s trying to get to the lower cost and scale that a service provider would get to. What do you think IT executives don’t get about private cloud? What aren’t they doing that would really take their data centers to the level that they need to be?

Manca: The technology, cloud technology, isn’t really that new. In a lot of cases it’s a repackaging of the old hosted models and they embraced that model. But that model was expensive and so they get a little nervous. There were a lot of these big outsourcing projects that you heard folks bringing back in because it was just so expensive. Cloud solves that problem, right? That’s really the problem that cloud is solving. The economic model is much, much better. But you’ve got to jump in. You can’t go halfway in with cloud. And I think that’s what I see, is folks want to build their own private cloud and they kind of ease their way down the path without going fully in and they don’t see the full benefits yet. The technology is quite solid. The ones I see that do jump in get tremendous benefits. Emory Healthcare is a customer that we’ve gone public with and they’re going to have up to a 25,000 VDI deployment on PAN Manager. They jumped in, fully embraced it, and they’re getting tremendous value.

Gallant: Is that to support kind of a bring-your-own-device or a much more intense focus on mobility?

Manca: It’s a mergers and acquisition thing, really. Emory is growing very fast by acquiring hospitals and small physician practices. You bring these folks in that have a myriad of devices that aren’t necessarily up to your corporate standard, but doctors don’t want to give up their favorite iPad or whatever their favorite tool is. You need to support that. And the easiest way for Emory to do that was to virtualize everything. Host it at Emory and use your own device and they virtualize it. It makes their management much simpler and it makes the end user’s life much simpler.

Gallant: What about the way you’re bringing this to market? What’s the overall marketing message?

Manca: It’s simplification. It’s reliability. We always have a high focus on high-availability disaster recovery because of the nature of our product. It’s openness. If you hit high availability DR and simplicity, the messages sometimes get confused across some of the other vendors that might have offerings as well. Openness is really a key differentiator. The other thing we can do is we can mix and match vendors. So if a customer wants to have a multi-vendor strategy or in the case of an Emory, they’re acquiring hospitals that may have different vendors in there, now we can put them under one umbrella and say, “Manage your HP next to your IBM next to your Dell. We don’t care. Put them in the same pool.” One of the things we’re seeing with service providers, which is interesting, is they may sell a different class of service based on the hardware you purchase. Maybe they see HP as an enterprise product and they see Dell as an SME, as an example. Maybe they can charge more for an HP server versus a Dell server. And so they can see differentiation in their own product offerings as a result of this capability.

Gallant: Last question: If you’ve got one minute with the top IT exec, what do you want them to know about Egenera?

Manca: We’ll simplify your data center. We’ll provide you an open solution. We’ll allow you to sleep at night by giving you high availability and disaster recovery capabilities. And we’ll give you new routes to market, we’ll give you new capabilities that allow you to bring new products to market much faster and much easier than you could in the past.

Read more about data center in Network World’s Data Center section.