Jeff Jaffe’s Blog

My last entry talked about the imperative to grow the application base for Linux via greater collaboration among vendors and standardization. In the future, I will have further comments on some of the more challenging aspects of getting this done. But before it gets too late, I wanted to comment on the other major excitement coming out of LinuxWorld San Francisco – progress on the Linux desktop.

Soon after I joined Novell, I started blogging about our technology directions. My first entry back in April 2006 was entitled “The Linux Desktop has arrived: The better desktop”. I argued that with SUSE Linux Enterprise Desktop 10 – due to ship that summer – that the time had come for more significant inroads for Linux into the desktop market.

It has been a year since we shipped this product. What has happened since? We have been building out our ecosystem, reaching a crescendo at LinuxWorld.

The first basic steps

Since we had built SUSE Linux Enterprise Desktop 10 to focus on enterprise deployment, it was critical to find demanding large customers. Also, to build the reputation required getting a referenceable customer. Our most visible relationship has been with PSA Peugot Citroen, with a major public endorsement of the Linux desktop.

In the enterprise, interoperability with Microsoft’s Office suite of products has always been a key requirement. So part of our November technology collaboration agreement has been to ensure the interoperability between Microsoft Office and Open Office. We have been building translators and making them available.

We are also interested in building alternatives for different client configurations. So in March we announced our thin client version, together with an imaging tool to allow customization to different client configurations.

The watershed event – five breakthrough partnerships at LinuxWorld

The basic steps above had intrinsic value, but also served to demonstrate to some of the largest I/T vendors that indeed it was time to embrace the Linux desktop. We were honored to capture five major resultant opportunities. With these landmark partnerships, we believe that the Linux desktop has reached the tipping point of mass acceptance.

IBM, Lotus’ Open Collaboration Client. At LinuxWorld, IBM felt that the time had come to fully embrace open collaboration by making a Lotus client available on the SUSE Linux Enterprise Desktop. We are delighted to have IBM’s endorsement of the Linux client and open collaboration. Novell’s customer base has enjoyed the benefits of open collaboration on Linux for some time, via the GroupWise and Evolution clients on Linux. Our customers have also expressed enthusiasm about Novell’s expansion of open collaboration through Novell Teaming + Conferencing, which will be available this fall. It is great to have IBM also providing support to open collaboration on the Linux desktop.

Lenovo Thinkpad pre-load. True acceptance of the Linux desktop is almost equivalent with the endorsement of a major PC manufacturer via a pre-load and support agreement. Exactly what we announced with Lenovo. This will move the market in a major way.

Dell. For anyone who attended Kevin Kettler’s keynote, you heard about the numerous places in which we are partnering with Dell, most notably in a preload distribution and support agreement with Dell in China. There is no question that the emerging markets represent a huge opportunity for the Linux desktop.

Tamil Nadu. We entered into an agreement with the Electronics Corporation of Tamil Nadu to equip 30000 student desktops with SLED.

Wyse. Wyse announced it it working with Novell around providing Linux thin clients. The announcement focused on easy imaging of thin clients with SUSE Linux Enterprise, as well as recent customer successes. More details will be forthcoming on this work soon.

The Linux desktop has truly arrived.

Those who attended Linux World heard Ron Hovsepian’s challenge to work for the continued growth of Linux. He called out three areas: enlarging the ISV ecosystem, enabling the next generation data center, and expanding the market (in terms of new offerings, partnerships, etc.).

In this posting, I would like to focus on the first of these – enlarging the ISV ecosystem.

Beyond the platform

We love Linux as an operating system. But in the marketplace, what makes an operating system sell is applications.

It is important to attract the anchor tenants – the mass market applications. Windows 3.0 was sold not on the basis of the operating system but on the tight integration with the Microsoft Office suite.

It is equally important to attract tens of thousands of additional applications. Niche applications. Line of business applications that are dedicated to specific industries. Single company applications built by the I/T shops of a company.

As a community, our larger competition is the Windows platform. Microsoft has many pluses and minuses. From an ISV perspective, Microsoft has one undeniable positive. If an ISV writes an application on Windows, it runs on Windows. Everywhere. (OK, almost everywhere.)

From a Linux perspective, if an ISV writes an application on Linux, it does not necessarily run everywhere. It might and it might not. Even if it runs on multiple Linux distributions, the market will not accept it unless it is certified on multiple distributions.

Attractiveness and Cost

How does an ISV look at this situation? It makes Linux a less attractive platform for building applications than it needs to be. Three reasons:

Confusion. As an ISV, I need to understand that there are many differences between platforms. I need to understand them, make choices where to certify, and keep track of the diversity.

Cost. Once I choose multiple Linux’s to certify to, I must go through the added cost to implement to different APIs, test to them, and certify to them.

Limitation of Opportunity. When I, as an ISV, limit my certifications, then I unnecessarily limit my opportunity. If I choose not to certify to a Linux which is popular in Europe, Asia, or Australia, I’ve cut myself out of a market.

These considerations affect different ISVs and different Linux distributions in different ways. For the small ISV, they will often de-select Linux distributions, limiting their cost and limiting their opportunity. For the large ISV, the anchor tenants mentioned above, they might certify multiple times, adding cost. They get the full opportunity of Linux, but deal with the cost and confusion.

These three issues also affect Linux distributions. The major effect:

Openness

A core value proposition of the free software community is openness. We share source code. We work together as a community because we want to remove barriers. We like choice.

By contrast, how do we appear to the ISV developer? The opposite of choice. We put forward the promise that everything is open. For the ISV developer, it looks very closed. The three problems indicated above: confusion, cost, and limitation of opportunity.

This lack of choice to the ISV developer reflects back to the Linux distribution itself, specifically the smaller ones. Large companies with significant Linux businesses such as Red Hat and Novell might have no problem attracting thousands of ISVs into our certification programs. But hundreds of smaller Linux distributions have no such capability. Where is our openness?

The solution

Ron’s keynote focused on the problem and less on the solution. We first need to agree on the characterization of the problem and the assertion that this is a problem worth fixing. Novell believes that we need to address it.

What might a solution look like? In Ron’s keynote, the key thought was to unify the Linux ecosystem.

We need to work more on standardization. We need to build upon the Linux Standards Base, but we need to take it much further. It must be done in a vendor-neutral way.

This will be a journey, not a quick fix which takes place overnight. Novell is committed to the journey, working with other key stakeholders of the Linux ecosystem.

Considerations

There are many dimensions to this problem:

• The framework to solve the problem. This includes assessing the difficulty of achieving this and building a staged technical plan.

• Analyzing how such unification improves our competition against non-Linux platforms; making “all boats rise” for Linux distributions.

• Benefits throughout the ecosystem if we reduce fragmentation – not only for ISVs, but for distributors as well.

• The tension between standardization (a good thing as mentioned above) and innovation (another core value that sometimes conflicts with standardization)

• Sacrifice. Analyzing where vendors (including Novell) need to sacrifice differentiation in order to unify the ecosystem.

• In a more unified ecosystem, where should vendors invest (less on differentiation; more on quality).

These considerations deserve a longer discussion and this posting is getting long. So we will defer that to a future posting.