Back in 1983, the Ethernet market was led by 3Com, Digital Equipment Corp., and possibly Intel. Cisco was still just a gleam in the eyes of Leonard Bosack and Sandy Lerner. TCP/IP was far from the leading protocol that used Ethernet as its underpinnings. Apple had AppleTalk, DEC had DECnet, Novell had IPX/SPX, 3Com had its own protocol.

You get the picture--networking was a mess. Multiprotocol routers were mostly a theory. Cisco jumped into this void with the promise that it would route just about any protocol that could be routed. Its hardware was robust, it performed well, and Cisco's engineering team would find a fix to practically any problem its customers had.

IT pros would like that Cisco back.

When we asked 588 IT pros (full report available at reports.informationweek.com next month) what they think about Cisco, and what they think the company needs to do to remain competitive, the top two responses were lower product prices and focus on your roots.

When Cisco was entirely focused on routing and later switching, it made great products that, through aggressive sales and pretty darn good customer support, kicked virtually every competitor to third-rate status. Those are some roots.

Cisco management tried to keep growing through diversification into everything from set-top boxes and Flip video cameras to telepresence, IP telephony, and now servers. That wasn't such a good move. Over the last five years, the company has watched its market capitalization drop and seen at least some of its customers question the value proposition of its products.

Cisco stuff is too expensive to buy, too costly to maintain, too complex to manage, and too proprietary--those are the common knocks against the company. Meanwhile the stock market demands fat profit margins and ever increasing revenue. It's a conundrum that CEO and chairman John Chambers and his team haven't found a way out of.

Chambers himself does just fine in our poll: 44% say he's either doing a great job or has stumbled in the past but is now on track. The next most common answer was "don't know" at 29%, followed by the 26% who want him gone. With a tenure as long as Chambers' that's not too bad.

Our survey also shows that despite all the grumbling, Cisco customers generally expect to remain so in the future. Our numbers show that 59% of respondents currently use more than 50% Cisco gear in their networks, and in two years, 55% expect to remain that way. While that downtick is subtle, more strikingly, 25% say they'll increase use of Cisco gear, 45% expect their use to remain the same, and 23% are looking to use less Cisco equipment.

You can certainly see those numbers from either a glass-half-full or half-empty point of view. We asked just those respondents looking to reduce use of Cisco gear why they wanted out. We offered 12 options, and a whopping 66% picked "prices are too high" among their top three reasons, followed by 36% who chose "product complexity" and 24% who thought Cisco products didn't interoperate well enough with others.

So, if not Cisco, whom do they like? We offered 12 options here and one big surprise popped up in the results. The top five included the usual suspects: Hewlett-Packard at 41%, Juniper at 40%, Dell at 24%, and Brocade at 17%. The shocker was VMware coming in a close third at 38%. Should Cisco worry about the software-defined data center rhetoric? You bet.

In free-form comments, by far the most common request was for Cisco to be more cost competitive. In reading the comments, you can almost feel the passion from loyal Cisco customers who are getting beat up by their CFOs and in some cases their CIOs to bring down the cost of networking. For most customers, Cisco can come late to software-defined networking, and it can probably even get away with some of its proprietary shenanigans. But if Cisco does both and remains the priciest option, even loyal customers will have no other choice than to look around. And increasingly VMware is giving them an interesting place to look.