Cisco And Deloitte Wrong: Good Practices More Impactful Than Vendor Choice

Cisco is fighting back against the notion that a multivendor network can simplify operations and reduce TCO. The networking giant commissioned a report by Deloitte that finds equipment operation costs will increase over the life of a network that uses equipment from disparate vendors. The fact is, one vendor versus multiple vendors is the wrong fight. Sound management practices and smart production selection will have a bigger impact on your IT costs than the number of vendors you use.

The spat started in 2010, when Gartner analysts Mark Fabbi and Debra Curtis wrote a brief, Debunking the Myth of the Single-Vendor Network, in which they singled out and then took issue with Cisco's stance promoting a single-vendor network "as a way to simplify operations, ensure reliability and lower the TCO for a network infrastructure." Curtis and Fabbi followed up in 2011 with The Disaggregation of the Enterprise Network, finding the "approach offers improved functionality and lower costs than a single-vendor architecture, with little or no additional operational complexity." Gartner concludes that going multivendor can save organizations between 21% and 26% in capital and maintenance contract costs during a five-year period.

The Deloitte report, Multi-vendor Network Architectures, TCO, and Operational Risk, is Cisco's answer to Gartner. Deloitte asserts that the initial cost savings cited by Gartner are mitigated over three years by "the incremental operating costs over the life of the equipment."

Deloitte's report contains an example where a company with a billion dollars in revenue has a $40 million IT budget, of which $8 million is dedicated to networking. The author estimates that the company could save 10% of the network budget, or $800,000, by going multivendor. That's 2% of the overall IT budget. Two percent isn't a lot by itself, but combined with other savings, it adds up. However, I bet the business case changes radically when your IT budget is much smaller because the cost of network equipment isn't exactly linear. Sure, smaller companies need less network gear, but the per-unit costs remains the same.

The Deloitte report states that the cost savings from a multivendor purchase can be gobbled up over three years in a couple of ways: increased training of new and existing staff and increased costs of a network management system required to run all the disparate systems. Let's look at those claims one at a time.

First, with regard to training IT to learn multiple systems, you'd still have to do that with a single vendor. Outside of switching and routing, Cisco has acquired all its other networking technology, as has Juniper. HP and IBM partner to fill out the gaps in their switching and routing products. None of these vendors offers a unified OS across all product lines (although Juniper is trying). So, it really doesn't matter if you buy from a single source or many--your staff still needs to learn to operate multiple products.

As for a single-pane-of-glass management, we haven't seen it yet and likely won't anytime soon. The combinations of vendors and products makes such a manger of mangers (MoM) a pipe dream whether we are talking single or multivendor models. At best, what we will likely see are management platforms such as those that are being developed in the cloud computing world where the system management is separate from the use of the system. In other words, network administrators design, monitor and manage the network using whichever management platform they choose; cloud administrators integrate with the network and use what is available within their private cloud software.The buyer's market itself also seems split between single-vendor and multivendor networks, according to our own survey data. We've asked our readers this question many times (IT Pro Ranking: Data Center Networks and IT Pro Ranking: LAN Equipment--free, one-time registration required), and the responses remain about 50-50 either way.

Jim Thomas, director of IT operations for Pella Windows, prefers a single-vendor approach. "If I buy all of the my network products from one vendor, the vendor has probably tested these components together in a lab, likely put the same set of products in the field and should have experience with that particular combination of products," he told me in an interview. "That means the products will likely work well together and, in the case of a problem, we'll get quicker resolution."

Greg Ferro, a network architect and designer based in the United Kingdom, takes a different tack. "Single vendor or multivendor is in the eye of the executive that thinks if he has one less vendor or supplier that his life would be simpler. The reality is that all of the product divisions within big companies are individual entities anyway. You aren't going to be dealing with a single person on a problem or a product--you will deal with the product specialist from each business unit. The problems and their solutions don't change when you have a single-vendor policy. It doesn't change anything at the operational level. "

Deliotte's report estimated that the cost savings of a multivendor approach could be wiped out by a single networking catastrophe. That's a silly scare tactic based on a set of cascading assumptions and dependencies. To be true, there would have to be a failure that is catastrophic enough to affect a large portion of revenue generating apps, and the failure would have to be due in whole or in part to a multivendor environment.

Finally, the resolution would have to made significantly more difficult due to the a multivendor environment. That is a whole lot of dependencies that have to fall into place to lay the blame on a multivendor network. Good engineering and operational management can mitigate the likelihood of an outage and reduce the severity and length should an outage occur.

There's no rule of thumb to go by. The decision to go single vendor or multivendor should rely solely on what you're trying to do and what fits best for your technical and business objectives. Neither a single-vendor nor a multivendor approach is inherently better or worse. You have to assess the benefits and drawback of each.

If you go single vendor, you get one account representative for sales and support, hopefully better buying power and the comfort of knowing that you have one throat to choke. However, you may also end up with inferior products that don't fit or perform as well as others. If you go multivendor, you may end up with more integration work and fewer discounts, but you are more likely to build an environment that is closely aligned with your objectives. However, if products from different vendors don't work well together, you have to lean on them.

Only you know the right answer for your organization.