Vendor lock-in possibly reached its zenith in the late 1990s and then began fade in part because giants like CA and IBM gained a reputation for charging hefty support fees that kept rising year over year for a customer base that, in many cases, had little option other than to keep paying. With mission-critical applications running on this hardware and software, it was way more expensive to migrate than to pony up.
The last decade, however, saw a marked shift away from lock-in. Propelled by the rise of Web-based architectures, Java, open source and far more powerful Intel/AMD processors (not to mention a lingering aftertaste from former years) the pendulum swung over to a multi-vendor approach. But that created its own issues such as complexity, lack of integration and constant maintenance/troubleshooting and heavier babysitting bill.
Could it be that the pendulum is once again swinging back?
“The more complex your technology gets, the greater the tendency to standardize and migrate to fewer vendors,” said Marc Barnett, senior manager for solutions and services marketing at CDW, a reseller of gear from a great many vendors.“When you mitigate all the variables, vendor lock-in simply makes interoperability and maintenance much more simplistic.”
True, agreed Jonathan Levine, COO of online marketer LinkShare. He used to field a primary application hosted on a large number of small Intel boxes running separate operating systems, analytics and reporting tools. Then he migrated to one large Oracle Exadata database machine designed to integrate the operating system, applications, database and hardware. The result? A much lower footprint, higher performance and a sharp drop in power and cooling demands.
“Historically, we have been best-of-breed, but the cause of downtime always seemed to be someone else’s problem,” said Levine. “We couldn’t live through another round of vendor finger pointing during a system failure.”
Bandaged fingers
Being burned in the past seems a common reason for a switch to single source. Rick Hassman, director of corporate applications at window manufacturer Pella found it hard to keep all systems synchronized.
“A bad experience with lock-in and a desire to achieve complete integration were the driving factors behind our desire to go with one vendor,” he said. “It is better in terms of time and money to have one vendor as you can form a real partnership.”
Barnett reinforced the point that vendor lock-in can facilitate ease of management. If you buy all of Brand A’s products, he said, they will usually work together more easily than mixing brands and having to worry about interoperability.
“When there is any kind of disaster, that causes a data integration problem, and vendor lock-in gives you one brand to turn to or one person to call for a solution,” said Barnett.
That’s the old, tried and true argument but it does not work in all cases. Sticking to one vendor, he said, may not actually provide the best-of-breed from product to product within a given solution stack. That may result in loss of an ability to create a competitive environment, and as such, to drive down cost or maximize the value you might gain from mixing the best of several brands.
Often with lock-in, you do not get a chance to bid out and earn the best pricing. It can also cause higher costs because it is harder to price out an integrated stack. Also, if the vendor fails to deliver on the promised support in a timely fashion, it’s your business, not theirs, that is on the line — at least in the short term.
“Working with a single vendor can dampen competition, innovation, and the openness that results from a more competitive environment,” said Barnett.
Case in point: Pitt Ohio is a trucking company that uses technology to gain an edge over rivals. As such, IT Manager Jules Thomas said that the organization favors a best-of-breed approach. It built a state of the art data center that is on display via a large window behind the reception desk. A potpourri of vendors such as Symantec, EMC, Data Domain, Apple, Cisco and HP all play a part. But even here, Thomas admits that he is taking a lock-in stance in one specialized area — moving to Cisco on the hardware side. That means Cisco switches for networking and storage, as well as Cisco server blades.
“Cisco put forward a compelling convergence and virtualization message via its Cisco Nexus 7000 core switches and the Cisco Unified Computing System (UCS) server blades,” said Thomas. “This will facilitate the move to 10 Gigabit Ethernet (10 GbE) throughout the enterprise for storage and networking.”
Unlocking networking
Just when it looked like Cisco was attaining dominance in networking and storage switch gear to achieve the ultimate in vendor lock-in, up steps an upstart to upset the applecart. In this case, open source networking.
Open source has a history of unlocking the marketplace in a variety of areas: Apache Web servers, Linux, storage and desktop productivity tools. Now, the networking arena is about to see more competition. Vyatta, for example, has introduced open-source routers and virtual routing appliances.
“Various markets are succumbing to the allure of open-system architectures,” said Kelly Herrell, CEO of Vyatta. “Vertically-integrated products that are 100% controlled by a single vendor eventually give way to solutions based on best-of-breed components.”
In other areas of the networking arena, the lock-in versus best-of-breed battle rages. Jeff Rountree, global network manager for Pump Solutions Group (PSG), monitors a data center in Redlands, Calif. and other sites in Michigan, Georgia, Germany, China, India, and France. He moved from using multiple network monitoring tools from a wide range of vendors, to centralization on SolarWinds Orion Network Performance Monitor. “Instead of getting tools from each vendor to monitor their own gear, we went with one that is more universal,” he said. “I can use it for just about any hardware or software.”
Scott Chapiewsky, senior consultant for Synamon Corporation, went the opposite way: He assembled an array of best-of-breed tools from SolarWinds, Plixer, HP and custom tools to manage the network used by the Denver Regional Transportation District (RTD), which provides bus, light rail, dial-a-ride and other transportation services for the greater Denver metropolitan area.
“Some of the tools are publicly available, while others are business specific tools to meet the needs of the organization,” said Chapiewsky.
Risk mitigation
Certainly, it can be risky to put all your eggs in one basket. Similarly, it can be unwise to have too many cooks spoiling one data center’s soup. There are advantages and challenges with each approach. “What it really comes down to is risk management,” said Barnett.
Whatever path is being considered, conduct a thorough risk assessment. Look at the pros and cons of staying as you are as compared to changing course.
Overall, Barnett seems to come out in favor of vendor lock-in … but only just.
“When you mitigate all the variables, vendor lock-in simply makes interoperability and maintenance more simplistic,” said Barnett. “The more complex your technology gets, the greater the tendency to standardize and migrate to fewer vendors.”
Gerry Gebel, an analyst with the Burton Group takes up a different slant to the risk equation. He said that if you are already heavily into using a certain vendor such as IBM, Microsoft, SAP or Oracle, it may be less risky to stay with that source. But whether one decides to go with a suite or best-of-breed approach, the most important consideration is integration — all the pieces need to work well with the other software and systems in place. This applies even when purchasing a suite, since that suite may be composed of pieces that were acquired by buying out other vendors.
“We hear from customers that they would prefer to deal with fewer vendors, which is why they go with a suite, but in many cases they do not live up to the promise of fully integrated technologies with few overlaps and few gaps,” said Gebel. “Even if you buy a suite, it may be like taking a best-of-breed approach from the viewpoint of deploying the products and operating them.”
The moral of the story, then, seems to not take the advantages of best-of-breed and lock-in at face value. A deeper view is always going to highlight nuances that are vital to the decision-making process.
Greg Schulz, an analyst with StorageIO Group cautions CIOs not to be too rigid in their stance — neither committed to best-of-breed nor over-zealous about vendor lock-in. There are pluses and minuses, benefits and nightmares to be had by each depending on circumstances. Those tending to one or other extreme are advised to take a fresh look and expand their options.
“The positive side of lock-in is that it enables an IT organization to focus on the delivery of their principal business and services while deriving maximum value from a specific vendor’s capabilities and without having to incur a business or financial hardship,” he said. “In other words, you are still in control of the vendor even if they have some lock-in.”
The downside, though, is when an IT organization achieves some value but becomes subservient to a vendor who becomes less than helpful or unwilling to remain competitive. Schulz said this shows up in increased costs and lack of long-term value for money from the technology being utilized.
And then there is the ugly side — paying a premium price for services and technologies from a provider who has become uncompetitive or unwilling to compete. Essentially, they now control the user, which, in turn, causes business bottlenecks or constraints to the organization.
“Keep in mind the Golden Rule: Whoever controls the lock-in or technology that could be a point of lock-in, controls the gold,” said Schulz.
Drew Robb is a freelance writer based in Los Angeles specializing in technology and engineering. He has a degree in Geology/Geography from the University of Strathclyde in Scotland. He is the author of Server Disk Management in a Windows Environment, as well as hundreds of magazine articles.