There’s a big debate brewing over bimodal IT.
Gartner says it’s essential to have two separate but coherent styles of work—one focused on predictability and the other on exploration.
IDC says it’s a recipe for disaster. In its just-published FutureScape: Worldwide CIO Agenda 2017 Predictions, IDC theorizes that by 2019, 80 percent of bimodal IT organizations will accumulate a crippling technical debt, resulting in spiraling complexity, costs and lost credibility.
Meanwhile, Forrester has weighed in saying there’s only one mode for IT operations—and that’s fast.
+ Also on Network World: IT spending on ‘innovation’ is now a priority +
So, as the industry analyst behemoths battle it out over the best strategy, I’d like to step back and offer a simple take on the bi-modal question, one that addresses the two seemingly incongruent points of view: Just how do you stay on budget and focus on innovation?
We’re all keenly aware that companies must continually look for ways to cut costs without ignoring the pleas from the IT department to invest in leading-edge technologies. The pace of digital transformations is quickening despite reminders that IT budgets will remain flat in 2017. What’s the best way to balance the need to explore new areas of innovation while keeping costs in check by maintaining a predictable, stable environment? It’s quite simple: Bifurcate your IT infrastructure and handle different parts of it differently.
Devising a bi-modal IT management strategy doesn’t pit two camps of IT professionals against each other, but rather categorizes and prioritizes how various types of IT assets should be maintained, upgraded and replaced. It’s a very practical take on the 80/20 rule.
Consider this: 80 percent of your IT infrastructure will last much longer than you’re being told, so focus on the 20 percent where you can achieve the most benefits in performance and innovation. Despite Cisco pushing you to upgrade network switches and routers about every five years, the Mean Time Between Failure (MTBF) on a Catalyst 3750 Series switch is 21.5 years and a Nexus 7000 tops 30 years, according to the company. Those kinds of assets typically make up 80 percent of your infrastructure, so replacing the workhorses prematurely only adds undue cost. Upgrading also opens you to unnecessary risk because new equipment can come with buggy software and take time to work out the kinks.
Equipment upgrades: cautionary tale
By now, we’ve all heard the sad tale of Peak Hosting, which was forced into Chapter 11 bankruptcy after frequent outages, exacerbated by buggy Cisco software, led its largest customer, Machine Zone, to drop them. To support a $48 million annual contract with Machine Zone, Peak bought $35 million of Cisco’s latest Nexus 3000 switches. In hindsight, the web hosting company probably could have sustained the business with the projected $6 million in hardware already in place.
Unfortunately, the new switches weren’t ready for primetime, and unstable performance took down Machine Zone’s Game of War for two hours. Now, both sides are warring over $100 million in lawsuits and counterclaims. This is a classic case of where opting for the latest and greatest over time-tested and field-proven switches wasn’t worth the risk.
If you have existing equipment that is perfectly adequate for the functions at hand, why not extend their useful life and redirect that budget to other areas? It makes good fiscal sense and keeps the door open to continuing innovations in areas that deliver the greatest business value.
Divide the network into 3 parts
No matter the size of your network, break it down into three parts—access network, data center and network edge—and determine where you keep what works and where you deploy state-of-the-art solutions.
For instance, the access layer includes bulletproof switches that deliver gigabit to the desktop all day long and are capable of 4K video streaming when the need arises. So, why spend 60 percent more on the newest switches only to achieve a 3 percent boost in functionality? Instead, focus on keeping pace with wireless access points and controllers, which are evolving every 12 to 18 months, while bringing impressive capabilities with each new generation.
At the network edge, focus less on routers and switches and more on firewalls because network security is still the most critical asset at this layer and will be for the foreseeable future. In the data center, servers, storage switches and disk arrays are stable, solid performers, while new functionality in the data center core and top-of-rack switches can produce measurable performance advantages.
Every time you swap out one piece of network gear for another, you need to do more than just calculate capital cost. Rip-and-replace comes with a big learning curve. Configurations must be converted. There are lots of potentially disruptive issues that are not easy to forecast but will bite you—and your budget—in the end.
Remember, stability is crucial, more so than ever. So, while the industry pundits duke it out to determine whether bi-modal IT is disruptive or disastrous, I suggest sticking with what works for as long as you can while constantly looking for opportunities to innovate your technology so you can truly differentiate your business.