Jeff Jaffe’s Blog

My next several posts will relate to BrainShare. I want to summarize my keynote address and discuss several of our significant announcements.

My keynote address started by discussing Ron Hovsepian’s business vision for Novell – Novell harmoniously integrates mixed-I/T environments to allow people and technology to work as one. I asked: what does this mean for our products? Novell is a technology company – we deliver technology that helps our customers solve their business challenges. So when Ron says Novell’s corporate positioning is Making I/T Work as One, how do we achieve that technically and how does the technical vision transform our customers’ I/T environments?

To address this, about three months ago, we assembled a team of Novell visionaries: our Fellows, Distinguished Engineers, and market-focused thinkers. We asked them: what technical foundation provides the next revolution in our industry? By looking at several use cases (see below), we settled on the notion of agility.

Agility

If you talk to any CIO today, I/T infrastructure is anything but agile. Yet that is what the CIO wants. The CIO wants infinite flexibility to deploy I/T resources against business objectives. In a word, that is Novell’s vision for the future – the agile infrastructure. Sounds like an oxymoron – the word infrastructure sounds inflexible. Our vision is to fix it. Fix it for both the data center and the user’s desktop.

We have given our vision a code name – The Fossa Project. Why “Fossa”? If you “Google” for an agile animal you come up with the Fossa. This most agile creature – living quietly in the jungles of Madagascar with no natural predator – is our model of agility. We like the name – it sounds like Free and Open Source Software, with agility. But not every technology immediately attracts a community, so we invest in proprietary technology to fill the gap.

Maturity

I mentioned above that we have been developing these themes for three months. The project is both very mature and very early at the same time!

• It is very mature. As I will discuss when I go through the technology section, this work is built on an incredible amount of work already done by Novell, by standards organizations, and by the open source community.

• It is very early. As I will discuss when I go through the technology section, there is a substantial amount of work to be done before this is substantially complete.

Why would we unveil a project that arguably is very early? This is not exclusively a Novell project – it needs to be an industry project. We need to work with our key hardware, software and distribution partners. Most important, we need to leverage the participation, innovation, quality, speed and drive of the open source community. Key pieces are essential to be done in open source, although we see a role for proprietary development where we cannot attract a community. So we want to establish a dialog with our partners and stakeholders and develop this in the open.

Megatrends enable an agile infrastructure

Technology trends have evolved to a point that agility is possible. So there is a match between the needs of an agile infrastructure and what is available in the technology storehouse. The key megatrends that are enabling this are:

• High Capacity Computing. To execute with agility, high-bandwidth communications, powerful multicore computers and open source virtualization technologies allow workloads to be executed anywhere.

• To express the needs of agility, policy engines provide techniques to express preferences for how workloads get executed, corporate resources remain protected, and regulatory compliance is ensured.

• Orchestration. To automate agility, algorithms allow the optimization and provisioning of these workloads.

• Convergence of telcommunications and I/T – or Unified Communications (UC) – is setting the IT agenda – and an explosion in collaboration paradigms is enriching the business potential of collaboration software. We see evidence of this in voice, video, email, vmail, web, blogs, wikis, twitters, team workspaces, real-time conferencing and visual voice mail, and in popular sites and products that bring them together, such as Facebook,YouTube, MySpace and iPhone.

• Mobility. Users with powerful mobile devices are driving the use of the compute infrastructure as a collaboration infrastructure.

The Fossa Project vision

How do all these megatrends enable agile I/T? To a CIO, agile I/T is a simple concept. Whenever a compute task needs to be executed, the I/T infrastructure should find the “best” place to execute that task. And it should be easy to do so. Defining the word “best,” however, is where all of the magic lies. The compute infrastructure needs to accept different notions of what it means to be best – so a customer can use I/T assets to provide value to the business. The value will differ in time: it could be to improve ease of use, scaling, innovation or other purposes.

This is where Fossa comes into play. To be sure, we are early in the development so we don’t yet have the full architecture. However, I want to introduce the core constructs here.

It all starts when a user has a workload that he or she needs to run. He or she makes a request of the I/T infrastructure – “the cloud.” What happens when that request goes into the cloud?

• The user needs a policy language to describe how and where the workload should get executed.

• We need to be able to talk about the items we are manipulating, so you need identity enablement of the users, the applications, the storage, the processor, even the virtual machines!

• The host systems must be able to run the workload. That’s where virtualization comes into the picture.

• Linux is at the core of our virtualization vision. A p-Distro or thin Linux is just enough operating system to get the hardware running and to host virtual machines. Then we put the identity-enabled virtual machines, or v-Distros, on top of the p-Distros. The result? Workloads can be dynamically moved to run on any policy controlled hardware.

• It might be nice if we could automate this process. That’s where automation and provisioning of the workload – also known as orchestration – come in. This is done with algorithmic automation and policy, together with continuous monitoring and load balancing.

• After the workload is executed, we need to know how it was done. The infrastructure must certify back to the user or I/T organization that the workload was executed appropriately. That’s compliance.

• These can be standalone workloads or workloads that also involve working with another human. In other words, collaboration.

Our vision provides two key attributes: interoperability and usability. Interoperable heterogeneity is needed because the customer might require specific applications running on specific operating systems. The usability is via user interfaces such as visualization and a simple policy language where the customer expresses requests and the sophisticated infrastructure handles it. So we provide a sophisticated infrastructure – capable of a great deal. However, we don’t surface the complexity. We surface the right knobs for the customer, the policy expression, to enable the customer to have agility with simplicity.

More to come

I want to provide a deep dive on our roadmap for these seven areas, but this post is getting long. So we will need to wait for the next post.

My December 31st blog was about Novell’s strategy for 2008. The major themes were Linux and open source, and products that manage a mixed source environment. Interoperability is key. On January 14th I followed this with a deep dive elaborating on our virtualization strategy.

Aligned with this strategy we announced last month that we entered definitive agreement to acquire PlateSpin, Ltd. PlateSpin builds products that manage virtualization in the enterprise: both proprietary and open source virtualization approaches. Clearly, PlateSpin sits at the center of all aspects of Novell’s publicly articulated strategy. Here, I’ll explore some of the key aspects, further.

Function

In Novell’s technical strategy for 2008, I emphasized that customers are demanding products that manage virtual machines. Customers want flexibility and ease-of-use to configure their workloads. They need tools to morph physical machines workloads into virtual machines – and vice versa; as well as tools that move virtual machines to different systems. PlateSpin, with its rich collection of P2V (physical to virtual), P2P, V2P, and V2V tools satisfy that customer need, along with capabilities to capture and deploy system images in an infrastructure independent way (e.g. P2I and I2V).

But PlateSpin’s capabilities go beyond the creation of physical and virtual machines workloads. There is automation built into the process. Automation is required if virtualization is to scale to be ubiquitous for production workloads. With the automation comes optimization – to get the most of the I/T organization’s resources.

Beyond the basic creation and automation is the deeper analysis. PowerRecon has powerful facilities to monitor and profile workloads which ultimately can be used to help provide for a more stable data center. It also establishes cost models, and then monitors the costs based on these models (typically utilization of system resources).

More recently developed are PlateSpin’s innovative disaster recovery products. By maintaining frequently snapshotted images of physical or virtual machine workloads, I/T managers have a clear recovery point for their workloads – should a disaster take place.

Agnosticism

Novell is a passionate believer in open source methodologies and sees a primary role for our company to bridge the gap between open source and proprietary software. But we don’t achieve this by keeping our head in the sand and working only on open source. We solve the customer’s problem – and that problem is to get all of the solutions working together. That requires a full commitment to a multiplicity of technologies.

In virtualization, there are many technology choices: proprietary, pure open source, and solutions that are based on open source in some fashion. The management problem is to treat all of these as relevant and drive for interoperability. Not to make choices. PlateSpin has done a terrific job – they are a companion product of choice for solutions based on VMWare, Microsoft, Citrix, Virtual Iron and others. We intend to keep that role.

Ecosystem

Another point I made in Novell’s technical strategy for 2008 is Novell’s focus on ecosystem. When customers are looking for I/T vendors to solve problems in interoperability, they are looking for more than agreement on standards specifications. They are looking for deep commitments among vendors that by working together and providing joint support solutions that we remove interoperability burdens from the customer.

PlateSpin’s agnosticism translates directly into a rich set of relationships with vendors: virtualization vendors, ISVs and hardware vendors – to ensure that customers have the best virtualization experience. We intend to keep that role. It fits PlateSpin and it fits Novell.

Synergy

While Novell endorses PlateSpin’s current direction and will keep that engine going, we also see opportunities to work together to take virtualization to the next level.

Several times in the past year or so (see here, here and here) I have articulated the importance of virtualization as a technology for Novell, and our participation in bringing an open source virtualization technology – Xen – to the market. Novell was the first to bring Xen to market as part of a Linux offering and more recently are bringing substantial workloads to Xen via our OES 2 product.

The full lifecycle experience for customers using virtualization needs to be outstanding. Customers require a well-integrated set of offerings: hypervisor, workload tools, provisioning tools, automation, optimization, etc. With tools from PlateSpin, our Open Platform Solutions (Linux) team and our Systems and Resource Management (ZENworks) team, Novell now has an opportunity and strategic intent to optimize the virtual data center and make it desirable to move workloads to our operating systems offerings that contain Xen.