Past Entry of Zippy's Telecom Blog
January 15, 2010
GoAhead Software’s Acquisition of S3’s embeddedMIND Business, and What it Means
GoAhead is known as a pioneer in the world of High Availability (HA) middleware. GoAhead focuses on providing five nines’ (99.999 percent uptime) availability via automation of the fault management cycle: fault detection, isolation, recovery, repair and health. This occurs through the key capabilities they deliver in their flagship product SAFfire which includes HA and management capabilities. SAFfire features strong support of the Service Availability Forum (SA Forum) open specifications — the Application Interface Specification (AIS) and the Hardware Platform Interface (HPI). The product is focused on carrier-grade and mission critical systems and supports detection of a failure and automatic policy-based recovery—typically occurring in a 500 millisecond (or so) window—resulting a stable failover unnoticed by the users.
And now, GoAhead enters a new, though related area—the configuration management space—by acquiring the embeddedMIND business from S3.
Yours Truly sat down to discuss this with GoAhead’s Bill Yaman, vice president of sales and marketing, and David French, director of product management, who has come to GoAhead from the embeddedMIND acquisition.
Richard Grigonis: So tell me, how did this all come about?
Bill Yaman: GoAhead Software, as you know, has been a market leader for COTS high availability software and services, with 60,000 live deployments worldwide, with an emphasis on the telecommunications, media, and the aerospace and defense markets. We’ve been involved in everything from the IP Multimedia Subsystem [IMS] to base station controllers, media gateways, AEGIS weapon systems, the U.S. Navy Littoral Combat Ship [LCS], and of course we’ve been a leader in the development and implementation of standards, as evidenced by our founding and participation in the Service Availability Forum [SAF], which reinforces our leadership position in the High Availability [HA] space.
What GoAhead is announcing today is that we’re extending our leadership position into a new technology area, although one that is certainly an adjunct to what we do today: the management infrastructure for network devices. The key element of our announcement is that we are acquiring the embeddedMIND business unit from S3. If you’re not familiar with S3, it’s an Irish company focused on being a leading provider of products for the digital TVs telehealth, semiconductor intellectual property [IP] and design industries. embeddedMIND has been a business unit within S3 since 2004, and we believe that it makes more sense for it to be part of a company that is more focused on the embedded software space, as we will explore in detail shortly.
GoAhead is retaining not only the intellectual property associated with the business but also the full embeddedMIND team which is located in Wroclaw, Poland. We’ll be maintaining that facility and building it out over time to be a Polish offshoot of GoAhead Software.
As for some of the industry factors that drove our decision to buy S3’s embeddedMIND business, as I’ve said, embeddedMIND provides management infrastructure for network devices. Just as GoAhead has for years provided COTS-based infrastructure software for HA, embeddedMIND is designed for equipment manufacturers who need to rapidly add embedded configuration and monitoring functionality their carrier-grade applications. So it really is an additional capability that’s needed for any networked device, really, because you need network management to configure the capabilities of such equipment, and that’s what the embeddedMIND solution is all about. Through a generic protocol independent management layer it provides a choice of Northbound management interfaces such as a command line interface [CLI], SNMP, Web/XML and Netconf, really a vision of providing a lot of flexibility to equipment manufacturers, who can now pick one or more of those protocols, and evolve and adapt and add others over time, all without having to reconfigure and re-architect their application.
What’s really cool about this acquisition for us at GoAhead Software is that embeddedMIND really sells into the same markets that we do, such as the telecommunications and media space, and they have a terrific customer base, including key Tier-1 equipment manufacturers such as Alcatel Lucent, Nokia Siemens and a bunch of others that are Tier-2 and Tier-3 players in the market.
Richard Grigonis: Were there any other factors that drove this acquisition?
Bill Yaman: In the telecom and media space, there’s been a movement away from proprietary systems. Organizations used to build their own hardware and software, operating systems and so forth. But then came the move toward utilizing the advantages inherent in adopting a COTS approach. Now we’re at a point in the history of the industry where all of the capabilities outside of the application itself that are needed in a networked piece of equipment really can be bought off the shelf. That’s true whether you’re talking about standard hardware such as ATCA [AdvancedTCA] or IBM Blade Center, or others; or whether you’re talking about standard OSes such as Wind River Systems or MontaVista, or GoAhead’s middleware, or embeddedMIND or Tail-f providing management infrastructure. Equipment manufacturers are really at a point where they can buy most of the capabilities they need off-the-shelf and they don’t have to spend the time and energy to build them themselves. Instead, they can now focus their time and energy on the value-added differentiation that really comes from the application itself.
However, I do believe there’s been a bit of a gap in the marketplace in that the equipment manufacturer at the end of the day by and large still has to do the integration of all these various components: hardware, OS, middleware and management infrastructure. We think there’s an opportunity in the marketplace to ‘pre-integrate’ those capabilities by acquiring technologies so we can take that burden off of the equipment makers themselves.
This evolution means that we as GoAhead can start to provide not just that single layer of the infrastructure, but multiple layers. By doing so we not only enable that pre-integration—which saves our customers a great deal of time and energy—but it also creates some interesting opportunities for what I like to call ‘1 + 1 = 3’ integration of those technologies.
So we’ll continue to work with OS vendors, hardware vendors such as RadiSys, Continuous Computing and others, but now we’ll start to provide a more complete set of system services capabilities, that being HA and the management infrastructure for customers who require one or both of those.
Richard Grigonis: Is embeddedMIND going to be integrated into your technology or is it going to be maintained as a separate, standalone environment? Will this easily dovetail with your other software?
Bill Yaman: Our vision on this is that embeddedMIND will continue to be sold as a standalone solution for those organizations that really need that management infrastructure. I know it’s hard to believe, but there are actually some organizations out there that may not need the HA capability. But really, any network device requires a network management infrastructure, and this provides us with a very broad set of market players to sell to who may not need HA but do need a standalone management infrastructure. So we’ll absolutely continue to sell it as a standalone solution.
However, we’re also starting to do work involving integrating our technologies. After all, if I’m buying a platform that needs to be highly available, the management infrastructure also needs to be highly available. So the first ‘no-brainer’ integration will be to make the embeddedMIND capability itself highly available. We’ll continue to do those sorts of integrations as well as some more integrations that we’ll reveal in the coming days and months. In any case, rest assured that we will provide both a standalone version as well as a version tightly integrated with our products for those customers who need those capabilities and don’t want to spend the time and money to integrate it all themselves.
David French: One additional point is that the market for embeddedMIND is actually wider than the market which GoAhead has accessed thus far. Typically, when customers are providing a new concept, then they make a small device, such as a single board, pizza box form factor, and they need to simply get it out into the market as quickly as possible. They always need management layer infrastructure and then when they switch up to a more scaled solution with multiple boards and a distributed architecture, then they start demanding a high availability infrastructure. That’s where GoAhead enters the picture. So this actually addresses the wider market and also makes the ‘startup scene’ more relevant to GoAhead’s customer base.
Richard Grigonis: It’s more of a cross-sell or upsell scenario.
David French: Exactly. The customer now has somewhere to go on their roadmap after they’ve got embeddedMIND on their system and they want to design and build a bigger ‘box’.
Bill Yaman: We have many customers who tell us that they need both capabilities, and they definitely don’t want to do the integration themselves. It’s all about time and money. They want the simplest thing that enables them get going the quickest so they can then bring their equipment to market faster than ever before.
The commercialization cycles are starting to shrink. The first step of that was that there are now COTS capabilities that allow equipment manufacturers to get products to market faster. There was a three-year window of conception to platform design and through field trials and marketing, but now that’s shrunk to more of a two-year window. These days, if an organization is left to integrate things themselves, there’s really just a two-year time frame to get the product from inception to market. What GoAhead is doing with our pre-integration is to take even more work out of that process so that we’re providing integration solutions that are quicker to get to market and hopefully are less expensive for manufacturers.
Solving Configuration and Management Conundrums
David French: Every device needs to be configured and monitored, so everyone typically has to deal with a subset of a command line interface, Web/XML, SNMP and NetConf. What differentiates the embeddedMIND product is that we’ve managed to achieve that capability in a generic fashion. Ironically, the main competitor of our embeddedMIND product is actually a company’s own in-house development team. The reason for that is that it’s quite difficult to make the management layer infrastructure in a way so that it can be reused from product to product to product. Even within the same company, it’s quite difficult to do that without hard-coding everything. That’s where our technology comes to the rescue.
The key piece to remember is that this is protocol-independent, so we don’t have an SNMP backbone in the way that SNMP Research have done in the past. Nor do we have a NetConf backbone in the way that Tail-f has with its product. In our case, it’s done in a completely programmatic way so that the customer is free to mix-and-match the northbound protocol that they need.
Configuration management is something that appears at various places in the device, depending on the architecture. Sometimes configuration is stored in the Management Layer, sometimes it’s stored at lower levels within the application. We support both of these approaches. There is a wide range of ways in which capability can be deployed. We help with the integration by providing an Integrated Development Environment [IDE], modeling support, and the ability to use new technologies such as the latest form of NetConf [Network Configuration Protocol] or you can use an Eclipse development environment system or you can code it up in source code if that’s your preference. These things can be mixed and matched, as usual. There’s also a wide range of databases, operating systems and applications in which it can be linked.
The key message here is that the whole package is flexible. We don’t come in trying to tell a customer, “This is the way your architecture is going to be.” Instead, we come in saying, “Yes, we can adapt to whatever your architecture happens to be.” I think that’s been a major factor in our success thus far.
Richard Grigonis: I guess the key to the whole thing is the additional software layer.
David French: Everyone has to build this software layer anyway. Usually when a company constructs such a thing in-house, it’s so hardcoded that they have to keep a specialized team onboard just to keep it alive. They usually need to rewrite it for any new platform or a new release of something else. The holy grail that everyone seeks is the easy re-use of code, and that’s where GoAhead come in. It’s not easy, but obviously it’s worth it.
The “pain point” as Bill pointed out, centers on integration. That’s where we see the big value in GoAhead, we’ve tackled it with our IDE and so on. Having a pre-integrated system with middleware is in fact unique in the marketplace at the moment.
Bringing it All Together
Bill Yaman: We support the notion of the application-ready platform. Moving from the ‘bottom up’ as always, we support a wide range of hardware platforms, mostly enabled through the SA Forum HPI specification that enables equipment manufacturers to have some freedom from being locked into a particular vendor, whether ATCA-based or otherwise. We have the HA middleware that we’ve been providing for a number of years now. On top of that, is the Application Interface Specification [AIS] through the SA Forum that, again, provides a layer of abstraction between the application and the middleware, protecting customers from ‘lock-in’ at that middleware layer.
And now the software layer we’re adding with embeddedMIND logically sits above the application. The embeddedMIND management framework provides a consistent protocol-dependent or protocol-agnostic management layer that enables an organization to plug in any number of Northbound interfaces, be it CLI, Web / XML, SNMP or the emerging NetConf standard. What we typically see here is that an organization may initially adopt one or two of these but then they want to enjoy the flexibility over time to bring one or more additional protocols into play as they mature, such as what’s happening today with NetConf. Thanks to our technology, these can be added without the need to re-architect the management framework layer.
Richard Grigonis: For developers, it sounds like it’s almost plug-and-play.
Bill Yaman: That’s where we need to get to. Again, I think we’ve made the transition so that COTS is now accepted, but there’s still a lot of work to put everything together. We need to ‘put it together’ to help equipment manufacturers so that components truly becomes become plug-and-play. COTS has reduced costs, but there’s still a big cost in integrating things, so we want to take that next ‘layer’ of cost out of the equation too.
In addition to overall product capabilities, one of our strategic directions here at GoAhead Software is to continue to focus on professional services that we can deploy at any point along the product evolution, from requirements identification through architecture, into development and on to deployment. One of the services that we deployed in mid-2009 that has been very well received by the market is an SA Forum Architectural Design Service, which at its simplest level can be said to help an equipment manufacturer design an SA Forum-compliant architecture that’s absolutely independent of SAFfire in any product, but our expertise in this space helps us help other organizations understand how to build and architecture that supportive of something akin to plug-and-play.
The combination of existing GoAhead technology and embeddedMIND are great. There are still great independent, standalone benefits of each of these that are very profound for the market, such as accelerated time to delivery, cost reduction and risk reduction, not to mention the IDE that David talked about that improves developer efficiency and cuts costs as well. But things get really exciting when you start thinking about GoAhead technology and embeddedMIND together. Now we have a single solution that a company can buy from GoAhead that provides a whole gamut of system services—not necessarily all of them. After all, there may be IP stacks or some arcane stuff in there, but certainly we offer more than has ever been offered from a single vendor. Our pre-integration concept further drives down costs and accelerates the time-to-market. A very simple example of this is the HA-enabled part of the pre-integration—it’s a no-brainer for us to HA-enable that management infrastructure, so that out-of-the-box the solution is as highly available as the hardware is.
Today, if someone were to buy GoAhead and play with a configuration/network management infrastructure, they’d have to do that integration themselves. We can now take that off of the table. It also provides us some opportunities for some interesting ‘1 + 1 = 3’ integration. I won’t tip our hand on that yet, but over the next 60 to 90 days you’ll see how we can bring these technologies together and, because they’re inside of the same company, some very interesting things can result that will have incremental value aside from having both of our technologies together on the same box.
Richard Grigonis: I imagine the word ‘synergy’ will be resurrected at that time!
Bill Yaman: Synergy is a good word for what’s happening. It’s also really important for our customer base that there is a single point of contact for support and services. When a customer has problems—and they will, because they’re writing software that’s never perfect right off the bat—they want to know that they have a single trusted vendor that they can go to that can resolve their problems and not have to worry. We now have a great model for this in the GoAhead services organization.
Richard Grigonis: Sounds great. I’ll be waiting for your next round of announcements.