Breaking the storage bottleneck

Ripley Hotch, Editor

'The focus in networking technology always shifts to the current bottleneck,' says Louis C. Cole, chairman of the board, president, and CEO of Legato Systems, Inc., a leading storage management company. Call it Cole's Theorem.

Today's bottleneck is storage. Legato's four founders anticipated this day more than a decade ago when they started the company, now with annualized sales of more than $120 million, with just $1.5 million in venture capital. The four visionaries didn't even have a business plan. Their experience, as the developers of NFS (Network File System) at Sun Microsystems, was enough to convince investors that these guys were onto something big when they foresaw the need for technologies that enhanced and simplified network computing.

Time has proven the founders, their investors, and Cole's Theorem right: Legato's share of the now $800 million storage management marketplace is already estimated at 15%. But the company expects the percentage to grow even as the storage management marketplace itself is projected to soar to more than $3 billion in the next three years. And this figure may well be an underestimate: witness the rapid development in storage area networks, network attached storage devices, and Fibre Channel protocols to support the bandwidth required for these technologies.

'What is driving the exponential growth in storage is the equally phenomenal growth in enterprise applications such as SAP, Oracle, Informix, and data warehousing that require huge amounts of storage capacity,' Cole says. 'With management costs running at six times the amount spent on storage itself, Legato is now in the right place at the right time.'

This wasn't always the case. In 1988, when the company was launched, there was no market at all for multiplatform, enterprise-wide storage management.

'Enterprise networking was just getting off the ground, most heavy-duty applications at that time were Unix-based, and virtually all backup products for Unix were based on the free utilities, dump and tar, that are still bundled with that operating system,' Cole says.

In fact, Legato's early days were less successful than expected. Although no one disputed the merits of even their first version of NetWorker, a Sun-based product that also supported DOS, the Unix crowd was not used to paying for utilities. Only marketing persistence allowed the company to break through this barrier, essentially creating an enterprise storage marketplace for the first time.

'The turning point for us,' Cole says, 'came when we convinced several Unix platform manufacturers, including Digital and Data General, to OEM the NetWorker source code. This gave us the confidence and incentive to continue our marketing efforts and, by the end of 1992, we had five OEM relationships going strong. But still we were only breaking even. Only in 1994 did we turn profitable when more and more of the Unix administrators who had tried our product began to realize how important quick and reliable restores were - and how much time they could save by empowering end users to do their own restores. The result: NetWorker started selling aggressively through the reseller channel.'

Cole came to Legato in its earliest days after leaving an early retirement. When he was 44, he sold his terminal emulation company, CXI, to Novell and then served as Novell's executive vice president of operations for a year at the request of Chairman Ray Noorda. That was enough of hard work, he decided, and he needed some time to play tennis, run, visit with family, and travel with his wife. So that's just what he did - for one year. 'During this time I did all the things I'd always wanted to do but put off because of my workload,' Cole says. 'It was a nice, casual, relaxing period.'

But then the lure of technology and entrepreneurial excitement became too great, so he started consulting for venture capital firms. During this period, he was also intrigued by a business plan for an ISDN telephone and organized some former associates together from his CXI days to start Combinet in September 1988.

Then, in mid-1989, he was recruited back to the technology workforce as president and CEO of Legato. 'I knew I was in good company with Legato's founders - Jon Kepecs, Rusty Sandberg, Joe Moran, and Bob Lyon,' Cole says. 'There was clearly a lot of talent, and, although development had started on NetWorker, they did not yet have a product. I thought this would be an interesting opportunity to start from the ground up and help these people succeed.'

Fortunately for the company, the storage management industry, and all players in that industry, this is where Cole has stayed for the last 10 years.

Legato's original plan was to develop an underlying architecture that would support a range of network enhancing services. Market research suggested that storage management would be a good first product, if it could meet certain objectives. First, it had to be robust enough for business-critical applications and to support the general complexities of networking. Second, it had to encompass multiple operating systems and platforms. As it turned out, the product would launch a revolutionary concept: end-user initiated, desktop-based recovers.

Serendipitously, while developing this product, Legato's founders also created a board-level NFS accelerator. Called PrestoServe, this product contains one megabyte of non-volatile, battery-backed-up memory and some software that does disk caching to enable it to temporarily store disk writes for subsequent uploading to the file system.

And it is with this product, not NetWorker, that Legato really launched itself. The reason: PrestoServe gave Cole the opportunity to exercise the now well-proven strategies that spawned the storage management market.

'I suggested we run with PrestoServe to gain some experience in marketing and selling a product and to create an early awareness of the company,' Cole says. 'Essentially, PrestoServe gave us the chance to establish the Legato name and to put the business processes in place that we would later need for NetWorker.'

Besides serving as a trial ground for learning how to take orders, ship product, and invoice and collect money, these business processes included the development of a reseller channel as well as the creation of OEM relationships. Some of these relationships survive to this day. In fact, every Digital server still ships with PrestoServe technology.

The reason Cole has always emphasized the importance of OEM relationships is his drive to be profitable. 'Achieving and maintaining profitability is always the most important thing to me,' he says. 'So we invest our resources where they will have the greatest impact based on current market conditions. And before we invest, we always do the research to make sure our investments have a chance to be successful.'

In addition to exploiting innovative marketing channels in this search for profitability, Legato has also pursued and driven new technologies. Legato was the first company, for example, to see the potential in automated tape handling.

'NetWorker was bundled with one of the first autoloaders developed in 1993, an 8mm tape device that looked and operated like a slide projector,' Cole says. 'Successfully marketed to oil companies, this innovation was noticed by a variety of technology companies, and the tape library industry was soon launched.'

Legato then saw the benefits of a completely automated storage management process and continued to adapt NetWorker to support virtually every robotics tape device in the marketplace. In so doing, Legato helped spur the growth of an industry that, in turn, has driven the need for automated storage management products.

This type of synergy has always served Legato, as well as the storage management industry. When research indicated that a more effective way was needed to protect databases, for example, Legato began working with Oracle to develop a standard interface that would allow not just NetWorker but all other storage management products to back up Oracle databases. This effort was successful, and Oracle ultimately created an open API for the entire industry in order to provide their customers with backup capabilities. And, in a testimonial to Legato's leadership position, Oracle also now bundles NetWorker with their database products.

Cole has also been quick to see the importance of storage area networks and network attached storage devices to both his company and to the storage management marketplace. A storage area network (SAN) is usually (though not necessarily) based on Fibre Channel rather than Ethernet, with storage devices (like disk arrays and tape libraries) instead of workstations. Servers would be part of the SAN, just as they would the LAN.

There are two immediate benefits to a SAN. The first is performance: Fibre Channel can run at 100 Mbps full duplex vs. the 20-40 Mbps that is common with SCSI today. The second is connectivity: in a typical storage topology until now, a given storage device would be associated with and addressable from one, or at most a few, servers. In a SAN, every storage device can potentially be addressed from every server.

'Our perspective on this new approach to storage is to evolve our product to support new market needs, and also to become actively involved in the development of storage area networking,' Cole says.

In fact, an argument could be made that Legato's efforts to create the multiplatform storage management paradigm have really been at the root of this entirely new trend. By demonstrating that disparate platforms and operating systems could all be protected through a single storage management server, the company essentially proved that data could be separated from the operating system used to create it.

Although Legato sees storage networks as the wave of the future, they do not believe that the current needs of the enterprise should be ignored. That's why they recently launched GEMS (Global Enterprise Management of Storage), a product that provides centralized control over distributed NetWorker servers. More than a storage management product, GEMS is really a Java-based framework architecture that can support a wide range of storage services.

'With GEMS,' Cole says, 'we're really getting back to Legato's original mission to create an infrastructure that supports a range of products to streamline network storage management. It is our intent to leverage this framework to support other third-party storage management solutions. In doing so, we will be able to extend our relationship with NetWorker customers, while empowering them with vital storage services.'

While saying that the first such enhancements should be expected this year, Cole remains elusive on details. 'We will take GEMS in many different directions, either through internal development, technology licensing arrangements, or acquisition,' he says.

What does seem certain is that the future is likely to repeat the past, with Legato setting the tone for the storage management industry. And, with Cole's Theorem guiding them, Legato is also confident of being ready for the next networking bottleneck - ahead of the competition.