Current Newsletter

ONF Summit Redux — The second Open Networking Foundation is coming up later this month at the Santa Clara Marriott (the first meeting blew out the obvious conference space at Stanford). I’m looking forward to seeing how things have evolved, although I seem to have been voted off the island for expressing some questions as to how all of this is going to actually penetrate the enterprise data center networking business (there is little doubt that large scale Web properties and the big network operators are going to use evolved forms of networking). The enterprise is a much more difficult propo­sition, at least until Cisco discloses its hand, because the practi­cality of a change hinges on how you repurpose all the existing equipment rather than unaffordable fork-lift replacement. In the last six months, I’ve been moving toward the “overlay network” model described by Arista and others in which the existing network evolves into a data fabric (what some call “L2 or L3”) and more of the higher level aspects of the network move into the hypervisor layer or virtual machine imple­men­tations. This discussion continues to emphasize the differences between the network and system commu­nities: the network people really don’t understand modern distributed system technology, and most system people are blissfully ignorant about how difficult it is to make a robust and reliable data fabric actually work. The overlay model is nice in that  regard in that it cleaves the network into two pieces, and lets the network experts continue to make the lower nasty bits work, and enables the system guys to build on that. The problem it poses for network vendors like Cisco is that more of the “network” value moves into software.  Peter’s Intro­duction to OpenFlow and Open Networking presen­tation can be found here.   — Peter Christy

Network Networking Dinner — Change in Date — Now Set for Tuesday, May 15^th in Palo Alto — You’re invited — How many newsletters do you get that feed not only your mind but your  stomach? After skipping a year our N-Square Dinner is on for May 15th. Doors open at 6PM for schmoozing and dinner is at 7. The topic for the evening is Social Business & Big Data.   Both of these topics are hot: How does a company integrate with and leverage all the modern forms of social networking? What’s realistic, what’s aspiration and what’s insanity? In our normal style we’re going to get a few key partic­ipants and stage an enter­taining and infor­mative debate. RSVP — John Katsaros

Intel and Apple — Intel just held their annual Analyst Summit, which is always a joy to attend because they are such an amazing execution business run by very smart people. A year ago, they were somewhat in denial about the shifts being driven by the iPhone, iPad and iCloud but they seem to have moved to the acceptance phase now. Intel is right to say that the PC isn’t going away anytime soon; they have every reason to expect robust low-end PC growth in developing countries as wages grow. But it’s striking to me how much the game has changed. A year ago they introduced the Ultra Book program aimed at getting very sleek, MacBook Air like notebooks into the market. The first ones introduced all seemed expensive, which is surprising considering that Apple has always been the “expensive” one. And then you realize that Apple consumes 50% of the world production in some flash categories, and has a clear supply advantage. Intel talks about how new systems like the Ultra Book can accelerate buyer refresh cycles and how much revenue that can bring into the PC ecology, which of course is very true until you compare it to (say) the iPad where the critics wonder if the new iPad is going to be sexy enough to cause buyers of the last generation to want one, not whether people will replace their five year old systems with a new laptop. I believe the cynics are right and that Apple at some point will come down from a trillion dollar valuation, but right now it’s Apple that’s trans­forming the world, driving by what people, like us, are buying. I held on to by beloved Blackberry until last fall, and didn’t buy an iPad until the 3rd generation. But I’m not looking back, not so much because the iDevices are sleek and cool, but rather because of the remarkable apps I can now use (and do). — Peter Christy

Apple and Security — Engineers in Silicon Valley don’t like Microsoft, by and large (in spite of Dan’l Lewin’s hard work for years), for under­standable cultural reasons. I don’t mind people having dislikes (I don’t like the Dallas Cowboys, as it turns out) except to the degree it colors their profes­sional judgment. For years, I have had to listen to engineers rant about how incom­petent Microsoft is, and how that explains why Windows is such an attack target, while I quietly held the belief (substan­tiated by data, for those who cared) that in fact Microsoft is remarkably competent in security, but as the dominating platform and appli­cation vendor, has the most surface area to protect. Perhaps this week’s very visible Mac virus will slightly balance the playing field. I don’t mind that engineers hate Microsoft if they want to (it’s well known that it’s easier to enjoy profes­sional football by disliking teams) but I worry that those  engineers will make bad product decisions based on a false under­standing of what Microsoft in fact does remarkably well.  — Peter Christy

“Getting It Right The First Time” —Thanks to many of our newsletter readers who bought the hard copy edition our publisher, Praeger, has issued a new paperback edition of our book. The big news here is that it’s priced at $19.95 on Amazon (and $17.96 for you Kindle owners). For those of you unfamiliar with our book, it’s about our method­ologies to accurately predict market conditions —  especially the market changes that will occur within the crucial 18-to-36-month innovation window.  Or, to paraphrase the advice hockey superstar Wayne Gretzky received from his  father: “Skate to where the puck is going to be, not to where it is.”  More»

Big Data…Big Infra­structure — When you list out the various market segments where Big Data plays a role one market that stands out is right under our collective noses — it’s log data  (and other forms of network, security alerts, server, enterprise app, storage systems activities) from the very devices that make up the infra­structure. Most people have seen Big Data appli­cations through the eyes of the high volume websites of the world — Facebook, LinkedIn, Yahoo!, Google… and the dream of these Web facing appli­cations is to improve ad targeting, keep visitors on the site longer, provide richer experiences… the stuff that makes serious money for the big Web properties. Meanwhile, others have been quietly sucking up infra­structure generated data sets with the intention of improving the opera­tional and cost aspects of their infra­structure. Splunk, the clear leader so far, filed their S1 (mandatory reading by the way) and is heading for a $1B valuation through an IPO which might happen any day now. Others in the space include Guavus, focused on complex service providers and Boundary focused on Big Data infra­structure. According to its S1 filing, Splunk has signed up over 3,000 customers in just a few years — tapping what is possibly the richest vein of the Big Data treasure hunt — optimizing infra­structure funda­mentals. Splunk is probably the best evidence that there is a “there, there” with regard to investing in Big Data startups.  — John Katsaros

Yet another anti-spam liquidity event — Eight years ago Symantec acquired BrightMail and started an anti-spam buying spree which continues to this day. Having started in 2002, ProofPoint was an early stage company when Symantec made its acqui­sition. Now, ten years after it started, ProofPoint is aiming for an IPO at a market valuation in the $400M range. We went way back in our archives to find our Anti-Spam Hall of Fame (below) — ProofPoint brings the total in the range of $3.5B. What’s remarkable about this is that when you ask experts who follow spam closely this question — “How many high volume spammers are there in the world” — the answer usually is: “Less than 100.” That’s right — 100 bad actors create the vast majority of spam and have caused a $3.5B business oppor­tunity. Funny thing — the worst thing that could happen to these companies is that the spam problem actually gets solved…but that’s another story. — John Katsaros

Sports and Big Data — I’ve been a sports analytics fan for a long time, especially when it came to Bill James’ efforts to create more meaningful baseball statistics. It turns out that MIT Sloan School holds a Sports Analytics Conference annually. For awhile I pondered attending it but in the end just watched the video highlight stream (most of the sessions are now up for Web video replay). It struck me that there is a lot to learn about Big Data from this community. In business, Big Data is often viewed somewhat unreal­is­tically in my opinion: you build a Big Data system and accumulate some inter­esting material, and then just hire a Data Scientist who will tell you how to run your business better. In sports this vision clearly doesn’t work. Data analysis in sports requires a lot of subject matter under­standing. In this conference, a great example is the presen­tation by ESPN on how they developed a new and more meaningful metric by which to assess quarterback performance. It took a lot of data and clever statistical analysis, but the more important element was creating the guiding hypotheses, the thoughts to be tested. A Data Scientist couldn’t do that unless he was also a football freak. I don’t think there is any doubt that profes­sional sports has a great deal to benefit from big data analysis,
but it’s also clear that most players and coaches aren’t very data literate (I think the same is true of business albeit perhaps less obvious). In sports, you clearly need to understand how to inject subject matter expertise into the process, and you clearly have to think a lot how to make use of the analysis. I bet these will turn out to be true in business as well. — Peter Christy

Big Week for Infra­structure IPO’s — Sometimes it’s painful to remember back in the 90’s when Infra­structure IPO’s were happening so fast that you couldn’t keep score. April 2012 is our “Back to the Future” moment as we get ready for three companies in their final stages for their IPO’s — S1’s are filed, preliminary prices are set — hopefully the Spanish housing debacle doesn’t sour the market. They’re almost there. Additionally, firewall maker Palo Alto Networks filed its S1 this month so it may soon follow. Our white paper “Big Traffic and the WAN” can be found here. — John Katsaros

Bubble Do Over — What would you do if you had the chance to relive the late 90’s and play the bubble for what it was? Well, it may be that we’re getting close to that stage again. Earlier this week Apple’s valuation broke the $600M marker just like Microsoft’s did at the end of the last bubble (today Microsoft’s market cap is $260B). Also this week there was the Facebook Instagram acqui­sition — $1B, 13 employees, $77M average per employee, 15 months old, 5M downloads. If that doesn’t bring you back to the late 90’s… — John Katsaros

TOP NEWS STORIES