How does one invest in software-defined networking (SDN)? This is a question I've been asked a number of times, and I must confess, it's a stumper.

I'm speaking for those who want to make public investments, of course, not people who want to put on khakis and a blue shirt and become venture capitalists. It's an interesting question because it's so difficult to answer. If somebody asks me how to invest in oil, I'll have an answer (Chevron), and I'll also have an answer if you ask me about storage technology (EMC).

Broad technology investment themes have been very useful over the years. For example, in the late eighties, if you figured out that "Wintel" was the developing duopoly theme, a dual investment in Microsoft and Intel would have done nicely for a decade. And the two companies are still chugging along and printing a lot of money.

The same thing with the Internet. No, investing at the height of Internet bubble mania wasn't a good idea, but when things settled out in 2001-2002, a nice slug of stock in Google, Priceline, and Amazon would have fluffed up the nest egg in the long run.

But SDN? Really, I'm stuck. SDN is hard to define. It's a broad term. It's software that bakes in networking capabilities. Also, it's a chicken and egg question: Are existing technology companies going to become more like SDN companies, or should we be looking at young companies that create a new SDN market?

And yet another question: Is SDN investment even a thing in the public markets?

I believe, in the end, it is. The big trend is that networking, and IT in general, is becoming more data and analytics driven. The cloud is SDN. It's the drive toward more automation of everything, whether it's your car or your Ethernet switch.

Thinking in these broad technological terms, we can get down to asking who will best execute in delivering a more software-defined environment.

The first company to come to mind is VMware. They are, unlike many of the other networking players, a pure software play. Therefore they are probably the best positioned at being a software-defined company. The drawbacks are that VMware's growth in server virtualization has slowed, and it's not a cheap stock — it trades at a price/earnings (P/E) ratio of 40 with a sub-20-percent growth rate. My longtime stock-picking formula says that ideally you should look with a company whose earnings and revenue growth rate are above the P/E. This is known in the business as the PEG ratio.

Fortunately, there's a trick here: You can still buy EMC, which owns a majority stake in VMware but itself trades at a 14 P/E. Yes, in the weird world of Wall Street you can buy the same thing in a different place at a fraction of the cost. Eventually investors will pressure EMC to do some financial engineering to monetize this situation. Expect it to divest its VMware stake, and its stock will rise.

What about coming at it from the other direction: hardware? Cisco is obviously a leading candidate to become an "SDN Stock," if you can look aside from the fact that it's a giant hardware company. Unlike VMware, Cisco trades at a more reasonable valuation of 13 times earnings, though it also suffers from slowing growth.

Cisco is probably worth a flier based on this idea: What if CEO John Chambers finally does leave? -- as I had incorrectly predicted this year.

The company, which has recently been great at piling up huge amounts of cash and not so good at growing, might eventually come under pressure from investors to spin out its higher-growth cloud businesses to unlock more value in cloud software. I think this is a realistic scenario, one in which Cisco's shares rapidly double. Cisco stock has been steadily rising recently, so I'm wondering if something is up.

On to some other candidates, the tech giants: Take a look at HP, Microsoft, and IBM. You could argue that there's some SDN software in there somewhere, though IBM has been backing out of the SDN market. These days, the "big techs," all kind of trade the same, though: They're big cash-generating machines with reasonable P/Es and lots of opportunity to "tap the next big market." Of the big techs, I like Microsoft and HP the best, with respective forward P/E ratios of 15 and 9.6. Coincidentally, they both have pretty good paths to become cloud-services and SDN tools companies.

Then there's Amazon and Google. Are they SDN companies? They're actually the biggest proof we have that SDN works and is coming. They've both built the largest private SDN networks in the world, even though they don't directly market SDN technology. The more you look inside, they show you what SDN produces: They take their data and use analytics from that data to tell them how to build a more efficient network to support their products.

Google is a no-brainer in any technology portfolio with a fully digestible forward P/E ratio of 17. Amazon is more of a reach, as the company's mysterious lack of profits and sky-high P/E (in the 300s!) is unappealing. But interestingly enough, the concept that investors have used to bid up Amazon shares — that it will become the software logistics and cloud provider to the world — is essentially the story of SDN.

So, I suppose this column is a bit of a cop-out, but it does get me to the bottom line: Right now, I would invest in SDN by picking a suite of reasonable technology firms that have some "exposure" to SDN. My list would be Cisco, Google, Microsoft, and EMC (as a proxy for VMware). I confess, it's a pretty boring list, but it will get the job done until the "BIG SDN COMPANY IPO" actually happens. These technology stocks are very reasonable at the moment — more reasonable than the valuations of technology companies have been historically.