This interview with Red Hat Vice President James Prasad, who heads up the company's embedded operations, discusses Red Hat's background in embedded systems, embedded Linux market strategy and product plans, and recent partnership with Wind River Systems. Prasad came to Red Hat via its acquisition of Cygnus Solutions in January, 2000.

---

Q: Will Red Hat sell embedded Linux solutions?

Wind River is our premier channel partner, much like IBM or Dell. Wind River is focused primarily in the embedded space. Our plan is leverage Wind River's expertise to best serve the customers.

Q: Will you sell direct?

Red Hat has been in the Embedded market and will continue to support our strategic customers. Wind River's sales and support infrastructure is set up specifically to serve the embedded market. Red Hat's focus has been on the enterprise side of business. We will leverage Wind River's infrastructure.

Q: Red Hat stopped marketing embedded Linux software products some time ago. Has it continued to do embedded Linux development on an engineering services basis?

Our strategy was to focus on the Global 500 customers in the Embedded market. In most instances, we were already working with these customers. Given this scenario, we had made the decision not to participate in market education and development. We've always been in the embedded Linux business.

Q: What kinds of projects have you been involved in, on the embedded side?

We have been involved in a broad range of applications. These included digital TVs, handheld scanners, and set-top boxes. In some cases we've provided complete software solutions, including the operating system and applications. In some cases we've just provided the OS. Our strategy has been to pursue a services based engagement in the embedded space. With the partnership with Wind River, we will expand this to include products targeting the Embedded market.

Q: Is Red Hat's interest in embedded part of an “end-to-end” enterprise strategy?

As we see devices becoming more convergent, and pervasive systems connected to systems that are much larger, embedded systems will play a larger role in the enterprise. This is one of the key points why we were interested in partnering with Wind River.

The Linux kernel lends itself well to end-to-end deployment. Other companies providing operating systems have to use different OS technologies for the different market segments. The base Linux technology can be deployed from high-end servers to desktops to small handheld devices.

If you look at the embedded marketplace, the overall size of embedded and enterprise are similar. However, the commercial market is much smaller in embedded. Our goal is by partnering with Wind River, we can promote and deliver standardized software to increase the commercial embedded market.

Q: The recent announcement of Red Hat's partnership with Wind River stated that you're working on a Red Hat embedded Linux distribution that will support x86 and PowerPC processors. What about support for ARM and other RISC processor architectures, which are increasingly popular in Linux-based devices?

Red Hat has been providing solutions for a broad range of processor architectures via custom to semi-custom solutions for customers in the Embedded market. Our initial focus with Wind River is to provide standardized products for PowerPC and x86 architectures. We will work with Wind River, our partners, and customers to expand our offerings.

Q: What role will eCos play in Red Hat's embedded strategy? Will the recent eCos copyright unification increase its use?

eCos was developed by Red Hat as an open source project. All the copyrights have since been assigned to the Free Software Foundation (FSF). The embedded market is changing. eCos is ideal for deeply embedded and low-end applications. As processors become more efficient, and memory is less of an issue, Linux is fast becoming the dominant embedded operating system. We have de-emphasized broad-based eCos support. We will continue to support our strategic customers. While Red Hat has de-emphasized eCos, there are other companies such as eCosCentric, that has some former Red Hat employees, supporting and advancing eCos.

Q: In the past, Red Hat has supported uClinux for resource-constrained MMU-less processors. Will Red Hat be including support for uClinux going forward?

Yes, Red Hat has supported uCLinux and will continue our support for strategic customers. Support for uCLinux is included in the 2.6 kernel.

Q: What embedded Linux products will you deliver? A kernel? Board support packages? Carrier Grade Linux? Vertical market products?

The initial target markets are the telecommunications, military, and aerospace vertical markets, on x86 and PowerPC architectures. We're working on a smaller footprint standard Linux distribution, based on approximately 100 RPMs, with special tools for optimizing memory footprint as well. Lower power consumption, smaller memory footprint, faster boot-up, and other features that are less important for these markets will be available over time.

Q: Red Hat had announced plans to produce a distribution compliant with the OSDL's Carrier Grade Linux (CGL) specification in December, 2002. Has it?

Red Hat has significant design wins in the telecommunications market where CGL features are key. We have shipped a general product that supports most of the 'priority one' features. Currently, over 90 percent of the 'priority one' features are supported by Red Hat Enterprise Linux, and we've had some good success. Glenayre and Tekelec resell Red Hat operating systems into the telecommunication market including one nationwide mobile carrier I can't identify.

Our customers understand our reasoning behind leaving some features out — some features that are and have been provided by other distributions. While they may be offering it now, the mainstream implementation may change as they make it into the mainstream, commercial distributions.

Over time our products will include more and more carrier grade features. Red Hat includes stable, mature, tested features. Bleeding edge features may break other functions important to markets beyond telecommunications. We're working with the OSDL to solidify the specifications and help support customer requirements.

Q: As far as having a Red Hat Embedded Linux product goes, rumor has it we're looking at two months from now?

We are investigating how quickly we can introduce Embedded Linux based on Red Hat Enterprise Linux (RHEL). I cannot be that specific on the schedule. We are aggressively working on introducing products quickly. We have engineering resources from Wind River supporting this effort. We are targeting RHEL 4 (Red Hat Enterprise Linux 4) for early 2005, for x86 and PowerPC support. We're also investigating whether we can come up with an embedded Linux product sooner, for x86, based on RHEL 3.

Q: How would you characterize the current state of embedded Linux, as a market and as a technology?

Linux for embedded is rapidly improving. If you look at power, footprint, fast bootup, and hard real-time, you will see greater improvements over time.

The biggest challenge you have in the embedded market is fragmentation and lack of standards that have led to inefficencies. This is a large reason why the commercial embedded market is small. Linux is being deployed in this market, but it's mostly roll-your-own versions of Linux that have replaced roll-your-own versions of other operating systems. This will continue to propagate inefficiencies. This will be the biggest challenge.

CELF and OSDL are great initiatives towards standardization. The goal of the combined efforts of Red Hat and Wind River is to help promote standards and binary compatibility so we minimize fragmentation and bring efficiencies into the embedded market.

Red Hat is the only truly open source enterprise Linux provider, and the same will be true for our embedded Linux. Open Source promotes standards. The acceptance process of modification and updates to the kernel by maintainers promotes standardization.

Q: What challenges do embedded Linux OS, tools, and services vendors face?

They face the same challenges as Linux, as a technology. Fragmentation has limited commercial market growth. If you can bring some standardization into the market, the market will be larger, and the commercial market will grow.

Q: What market opportunities do you see for Linux in the embedded devices and systems market?

An end-to-end deployment all based on the same base Linux technology is a huge opportunity. Linux lends itself well for use in small handheld devices, to large IBM 390 machines. What this means is that most of the new systems will be complex, sophisticated, intelligent devices that will be connected. Seamless operation amongst the systems will become very important.

As standards becomes more prevalent and fragmentation is minimized, there will be increased opportunities for third parties, ISVs, IHVs, etc. As the embedded market evolves, and efficiencies are introduced, players in this market will focus on their respective value-adds, enabling each to focus on their areas of expertise and provide more value to their customers.

Historically, Open Source has done well in markets where software technology is reaching some level of maturity. This is where projects such as Mozilla, Open Office, Linux, and GNU are doing well. What you will see are similar projects optimized for embedded getting stronger.

Q: What challenges does embedded Linux face as a technology?

Lowering total cost of ownership is one of the reasons open source is so attractive. Users of this technology pay only for the incremental investment required to make the technology useful for them. In many instances this is an order of magnitude savings, largely because someone else has invested in moving technology forward. This also has the effect of a much shorter time to market advantage. For this reason, some may delay contributions, or not contribute back at all. The value of open source comes from technology being contributed back. The model works best if everyone contributes and we all compete on our value-add to our respective customers.

Q: What embedded Linux technology developments do you find exciting?

Recent developments are all very exciting. The 2.6 kernel, and different groups working together to drive standards, such as the OSDL and CELF. Red Hat is about standardization and getting efficiencies into the marketplace. We welcome these.

If you look at the proliferation and adoption of Linux in the enterprise, that's very exciting. For the first time, during Red Hat's last fiscal quarter, there were more new Red Hat Enterprise Linux subscriptions then Solaris subscriptions during the same period.

Q: How do you see the Eclipse tools platform fitting into the future of embedded Linux development?

Red Hat is very actively involved with the Eclipse Foundation. I believe Eclipse will become the defacto standard for integrated development environments (IDEs). Wind River's Workbench IDE is based on Eclipse. If they have not already done so, many other IDE vendors will be introducing IDEs based on Eclipse technology. Many of the development engineers are using Linux as their preferred host.

Q: How do you see the 2.6 kernel affecting the embedded Linux market?

RHEL 3 already supports many 2.6 kernel features. It includes the Native POSIX Threads Library (NPTL) and O(1) scheduler that are significant for the embedded market. RHEL 3 is the only commercial Linux that has back-ported 2.6 features. Kernel pre-emption, uClinux, kernel configuration options for removing what is not required for embedded applications, and the new scheduler in 2.6 will all be benefits.

Q. What's your vision of the future for embedded Linux? That is, how big will the embedded Linux market get, and when will it start to reach that peak?

Analysts have previously forecast larger markets, but taken them down significantly. This is typical where new markets emerged based on new technology. There has been strong negatives effects due to lack of standards and fragmentation in the market. The market could be much larger. As efficiencies are introduced in the market, as intelligent devices are all connected into the Enterprise infrastructure, the market grow significantly.

Q: Can you share one or two of your company's most exciting successes?

We've established a financially viable, sustainable business model, a subscription-based business model. We have also established an Enterprise product delivery model based on standard, binary compatible products that have a longer product life cycle and longer support. This is completed with training, services, and support.

It is exciting to have established a profitable company that is based on the open source technology!

Q: How about a failure?

The paths to get to financial profitability have not always been attractive. It's been a learning experience.

---

About the interviewee — James Prasad has over 20 years experience in the high technology industry. His broad experience covers engineering, sales, and marketing roles within start-ups as well as multi-billion dollar organizations. He has an undergraduate degree in engineering with an MBA in marketing. Prasad joined Red Hat from Cygnus, following its January 2000 acquisition, and was appointed Director of Business Development for Embedded technologies. He was appointed Vice President and General Manager of Red Hat EMEA (Europe, Middle East, and Africa) in September 2001. He is now Vice President of Global Engineering Services and manages Red Hat's Embedded business.

---

 
This article was originally published on LinuxDevices.com and has been donated to the open source community by QuinStreet Inc. Please visit LinuxToday.com for up-to-date news and articles about Linux and open source.