VenChar

The coming DNS challenge

Om links to a Business 2.0 article on the dangers of a over-worked DNS and how that has resulted in the recent outages at Google and Comcast.  The article and the resultant press coverage on this topic highlights the growing problem with current DNS infrastructure. 

(Warning: tooting Globespan and our portfolio company's horn ;-))

We had invested in Nominum in early 2003 on the investment thesis that open source BIND will not scale and its nice to see that atleast once in a while we are right with our investment thesis.  Nonimum has assembled the world's top experts in DNS (Paul Mockapetris, David Conrad, etc.) and they have built the most scalable DNS engines ever.  If you are a carrier, ISP or an enterprise, you would be doing yourself a disservice by not upgrading your DNS architecture.  If they can't find you, they cannot buy from you.

(end tooting)

Update 1: Good comment by Jeff Nolan on the comcast outage:

that comcast outage wasn't due to overworked DNS but rather a move that the company made to consolidate their regional DNS servers into centralized data centers. Comcast screwed up their network through poor planning, and as evidence to prove my point that this wasn't a DNS scaling issue is that comcast users en masse started moving to Verizon's DNS servers (I did) and they handled the load just fine despite having more broadband customers than Comcast has.

I think there is a fair debate to be had about how to scale DNS, especially when we consider exponential scaling for VoIP to the masses and so on, but the other plane that this debate should reside on is distributed vs. centralized.

I don't know what happenned at Comcast and Jeff is usually right on these kind of things so I am sure it was due to poor planning.  What I do know given our active involvement in Nominum is that the DNS infrastructure is being taxed in numerous ways:

• DDOS attacks are dramaticallly impacting DNS availablity and BIND is not set up to handle that
• Pharming attacks are raising issues of what "DNS addresses" to trust and how to know you really are going to the right address
• Network operations are becoming complex and network operators want to move, centralize and manage their IP addresses and the current bubble gum and wire setup (usually a bunch of excel spreadsheets) don't let them do that.

Given all that, I think more and more enterprises and carriers are going to opt for a commercial DNS solution rather than rely on BIND

Detecting nukes in transit: What can the newly-established DNDO do?

Just finished writing a paper with Sri and Tom Tisch - it's titled 'Nuclear Detection: Portals, fixed detectors, and NEST teams won't work on a national scale, so what's next?'. We analyze the *use* of nuclear detectors to help prevent terrorist nuclear attacks, and we conclude that fixed detector approaches (such as those currently being implemented) are unlikely to be that effective. Here's the executive summary of the paper:

Recognizing the need for detecting terrorist attempts to transport or use fissile nuclear materials, President Bush’s FY 2006 budget request includes $246 million to form a Domestic Nuclear Detection Office (DNDO) within the Department of Homeland Security (DHS). [1] “The DNDO will provide a single accountable organization with dedicated responsibilities to develop the global nuclear detection architecture, and acquire, and support the deployment of the domestic detection system…” [2] How can DNDO planners deliver a global nuclear detection architecture that works?

Nuclear detection systems, as architected and deployed today, leave loopholes in the transportation network that terrorists can easily exploit by making use of light road vehicles to private jets to oil tankers [3].  Progress can be made if we face up to three fundamental facts:

1. Terrorists will most likely try to use highly enriched uranium (HEU), not plutonium: assembly of a HEU bomb does not involve technically complex detonation as with a plutonium bomb.

2. Terrorists can circumvent a network of fixed detectors: fixed detectors not only lack sufficient proximity and exposure to the vehicle in transit but also do not screen many types of vehicles.

3. R&D breakthroughs cannot change the physics of detection: passive detection of HEU will always be limited by its natural rate of radioactivity, and the attenuation of radioactivity is very sharp with distance [4]. The gamma rays and neutrons useful for detecting shielded HEU permit detection only at short distances (2-4 feet or less) and require that there is sufficient time to count a sufficient number of particles (several minutes to hours).

Recommendation: Due to fundamental physical limits, the current trend toward a fixed detector infrastructure is a dead-end. The only way shielded HEU can be effectively detected is if commercially-available detector technology, rather than being kept at fixed locations, are directly integrated into vehicles themselves. Detectors would travel with vehicles and have enough time to record radioactivity before reporting their readings to a network of check-points (in the same way E-Z pass collects highway tolls).

Our paper, 'Nuclear Detection: Portals, fixed detectors, and NEST teams won't work on a national scale, so what's next?' explores tradeoffs in detecting HEU in transit, and analyzes its technical, operational, and economic feasibility.

---

[1] “R&D in the Department of Homeland Security”, AAAS, http://www.aaas.org/spp/rd/06pch12.htm

[2] “Fact Sheet: Domestic Nuclear Detection Office,” http://www.dhs.gov/dhspublic/display?content=4474

[3] Medalia, J., 2005, “Nuclear Terrorism: A Brief Review of Threats and Responses,” CRS Report for Congress, The Library of Congress http://fpc.state.gov/documents/organization/43399.pdf

[4] attenuation of radioactivity with distance is subject to an inverse-square law in free-space and is exponential with shielding

Portable People Meter

Interesting NYT article on Nielsen and Arbitron and the changes brewing in TV viewership tracking.

For the past few months, Arbitron has been taking a distinctly unorthodox approach to measuring audiences. Currently the company is recruiting a couple of thousand volunteers in Houston and asking these randomly chosen men, women and children to wear a black plastic box that looks like a pager, three inches by two inches by one-half inch, whose circuitry is roughly as complex as that of a cellphone. In the radio and television industry, this little box is known as the portable people meter, or the P.P.M. In both a business and a cultural sense, it also seems to be the equivalent of a large explosive.

The Houston volunteers will clip the P.P.M.to their belts, or to any other article of clothing, and wear it all their waking hours. Before going to bed, the volunteers will be expected to dock the P.P.M. in a cradle so that overnight it can automatically send its data to a computer center in Maryland, where statisticians can download and review the information. There are still kinks to work out, but ideally the P.P.M. will tell Arbitron exactly what kind -- and exactly how much -- television and radio programming a person was exposed to during the day. Eventually the P.P.M. may also tell the technicians at Arbitron a host of other things too, like whether a P.P.M.-wearer heard any Web streaming, or supermarket Muzak, or any electronic media with audible sound that someone might encounter on a typical day.

The technology underlying the PPM (currently trialing in Houston) is interesting: they bury a unique, repeating, inaudible digital code in the audio tracks of TV and radio shows. The PPM picks that up. Arbitron is asking radio and TV stations to embed this code in their programming.

To date, viewership tracking has been very unscientific and potentially highly inaccurate. Techniques like this may bring rigor to the data-gathering process and allow the old media to compete more effectively with the web. They might also, in the end, lead to a shift in how and where advertisers decide to spend their money. The underlying conceit, as articulated by Arbitron's CEO, is this:

''Media is following you not just when you consciously turn on your satellite radio in your car, or when you consciously flip open your cellphone and get some cable channel delivered to it,'' Morris told me. ''It's also coming at you when you walk through Grand Central station. It's on the floor and on the walls. It's coming at you at the malls, where the L.E.D. screens are all around you along with the piped-in music. Advertising is becoming incredibly ubiquitous, so you need measurement that is equally ubiquitous.''

The PPM can do things like register the impressions on ads that screen at movie theaters; with a GPS-additive it could track whether you walked past a particular billboard; with RFID capability built in, it could tell that you happened to pick up a copy of the New Yorker, etc. etc. It's a short step from there to tracking the correlation between ad impressions and buying behavior.

Finally, here are some numbers:

• Nielsen does about $700MM in revenues
• Nielsen's People Meter (not the experimental PPM) is in 8000 homes
• TV advertising is at $60B annually
• On a typical weeknight, about 100MM Americans watch prime-time TV
• The average household watches 8 hours of TV per day

Millimeter-wave sensors

The NYT has an article on millimeter-wave imaging technologies applied to the detection of concealed weapons. The human body has a high emissivity and emits a great deal of millimeter-wave energy (between 30 and 300 GHz)- it shows up as hot on a millimeter imaging system. By contrast, a concealed gun, for instance, has a low emissivity and a high reflectivity - it reflects the ambient energy (at the temperature of the surroundings) and shows up as cold on the scan. The temperature differential with respect to the surroundings allows for the discrimination of the weapon being carried.

The article profiles three companies (Brijot Imaging Systems, Millivision Technologies and Trex Enterprises) that appear to have working imaging systems integrated with video surveillance and software that accomplishes detection and classification in a device that's about $50K.

Interestingly, these passive detection systems have an active counterpart (involving bouncing millimeter waves off the subject in a manner analogous to radar). Understandably, there are health and privacy concerns around the active imaging systems as a result of which the passive systems are likely to get better traction.

Millivision has a nice whitepaper on the technology on their website.

Radiation detectors on buoys

The Lawrence Livermore National Labs site has an interesting write-up on trials of radiation detectors aboard buoys off the coast. The idea is to detect nuclear materials that might be carried on board boats and other vessels before they get close enough to land to be dangerous. The detectors are powered by wind- and solar-powered generators and are outfitted with wireless communications links.

Homeland security experts are evaluating a wide range of possible threats from terrorists. One of the more troubling scenarios is a small and crude nuclear device transported in and detonated from a boat located near a naval military base or a civilian shipping terminal. Thanks to a Livermore design, buoys outfitted with commercially available radiation detectors could soon play an important role by warning of the presence of nuclear materials in marine environments.

9/11 showed us that we needed to secure civilian transportation modalities (a shift away from the cold-war thinking of building missile shields, etc.). If the trials are successful, these detector systems might be deployed around busy ports to interdict and deter marine transport of nuclear materials and weapons. Apparently, proposals have already been submitted to deploy buoys with radiation detectors in the Oakland harbor.

Curious to see what the specs are on the detector system: how well detection at a distance works, how high the false positive rate is and how closely the buoys need to be spaced in order to be effective. As with any RF system, radiation has a power-law falloff (inverse-square law in this instance) with distance...

Tire pressure sensors

The National Highway Traffic Safety Administration (NHTSA) (part of the Department of Transportation) is going to mandate the installation of tire pressure sensors in vehicles starting in model year 2006. If the tire pressure falls below the recommeded level by 25% or more, the car will alert the driver to the fact. The rulemaking is motivated by the fact that underinflated tires are the cause of several accidents: it is estimated that the rule will save 120 lives and prevent 8400 injuries annually.

According to NHTSA, under-inflated tires can adversely affect fuel economy, lead to skidding and loss of control and hydroplaning on wet surfaces. It can also increase stopping distance and the likelihood of tire failures.

It will cost the auto industry on the order of $50-$70 per vehicle and save consumers about $30 due to better fuel efficiency, fewer crashes and longer tire life. According to the NYT, "The government estimates that it will cost the industry between $800 million and $1.1 billion to phase in the technology on all new vehicles from this year through 2007."

Apparently, it all started with an act of Congress in 2000 (prompted by the Firestone recalls) requiring the NHTSA to set guidelines for tire pressure monitoring systems. The agency dragged its feet for a couple of years (apparently due to the lobbying efforts of the auto industry) until safety groups sued multiple times to get the process moving and received court orders directing the NHTSA to act quickly.

For more background see the NTHSA announcement and the rulemaking. Interesting story because it illustrates the dynamics of rule-making (conflicting interests of safety groups, Congress and the auto industry), the time-scales for the passage of rules in this industry (~4 years), the process for mainstreamization of new technology (tire-monitoring systems are already in place in 2-4 MM luxury vehicles today and the adoption of this technology will be accelerated by this ruling) and because it illustrates the reactive nature of such efforts (it took a well-publicized system failure to spur decisive action).

Talk radio

Read David Foster Wallace's entertaining dissection of the talk radio industry in the Atlantic on my plane ride back from Atlanta yesterday. Some excerpts:

Whatever the social effects of talk radio or the partisan agendas of certain hosts, it is a fallacy that political talk radio is motivated by ideology. It is not. Political talk radio is a business, and it is motivated by revenue. The conservativeness that dominates today's AM airwaves does so because it generates high Arbitron ratings, high ad rates, and maximum profits.

Radio has become a more lucrative business than most people know. Throughout most of the past decade, the industry's revenues have increased by more than 10% a year. The average cash-flow margin for major radio companies is 40 percent, compared with more like 15 percent for large TV networks; and the mean price paid for a radio station has gone from eight to more than thirteen times cash flow.

Also, here is an interesting story about a rather clever show:

... the Phil Hendrie Show which is actually a cruel and complicated kind of meta-talk radio. What happens every night on this program is that Phil Hendrie brings on some wildly offensive guest - a man who's leaving his wife because she's had a mastectomy, a Little League coach who advocates corporal punishment of players, etc. - and first-time or casual listeners will call in and argue with the guests and (not surprisingly) get very angry and upset. Except the whole thing's a put-on. The guests are fake, their different voices done by Hendrie with the aid of mike-processing and a first-rate board  op, and the show's real entertainment is the callers, who don't know it's all a gag - Hendrie's real auidience, which is in on the joke, enjoys hearing these callers get more and more outraged and sputtery as the "guests" yank their chain.

Later on in the piece, Wallace goes on to speculate that perhaps the callers are also fakes conjured up by Hendrie and his crew, so that the "initiates" are also the butt of the joke. Rather funny.

Also learnt about Arbitron which is something like Nielsen ratings for radio. It is referred to as "Arbitraryon" by insiders because "it is 100 percent diary-based, and diary surveys are notoriously iffy, since a lot of subjects neglect to fill out their diaries in real time (especially when they're listening as they drive), tending instead to wait till the night before they're due and theen trying to do them from memory." Leading, no doubt, to inaccurate ratings numbers that are fed to gullible advertisers.

Experimentation and Failure

Last weekend I attend the Stanford Entrepreneurial Conference.  The lead keynote speaker was Jeff Bezos of Amazon and he did a terrific job.  He talked about historical innovations and innovators.  Some of his examples were:

1. The guy who invented toilet paper
2. The woman who invented windshield wipers - prior to her invention people just stopped every mile or so and wiped the windshield
3. The woman who invented "white out" - she was an executive assistant who made a lot of mistakes while typing and was annoyed there was no quick way of correcting her mistakes.  By the way "white out" was nothing but white paint - just shows you don't have to always invent cutting edge stuff to be successful

Then he talked about Amazon and how he tries to instill a culture of "innovation' within Amazon.  He said the two important factors that help drive innovation at Amazon are:

• Cost of Experimentation is low - its easy to run experiments in an online business and check to see effectiveness.  Amazon runs experiments all the time and evaluates which of his innovations are more effective
• Cost of failure is low - again another terrific element of online businesses is that the cost of failure can be controlled since you can run these experiments without a significant cost if it fails.

The combination of these two factors also result in an open non-hierarchical organizational culture at Amazon.  Even front-line folks can run a experiment since the cost of doing and the cost of failure is low.  These two factors are not unique to Amazon, they apply to all online businesses and good online businesses take advantage of them - the folks at Plaxo (full disclosure:  portfolio company) are masters of this.  Rikk, Todd, and Cameron and the other engineering folks constantly experiment, test and iterate on their new feature set before releasing it.  No wonder they are at 5 million users. 

APIs and innovation

I had written a while back about Skype's API. I found out recently that someone's developed an answering machine application using the API:

SAM is a simple voice answering machine for Skype® users.

When you are away from your PC and there is no one to answer your incoming Skype calls, SAM will pick up the call, play a greeting message and the "all-time clasic beep" so that the calling party will leave a voice recorded message.

It looks a little rough around the edges right now, but I still think it's a great example of how opening up APIs, done well, can fuel user-driven innovation. Apropos of this, also see the recent Google AdWords API announcement:

Google's free AdWords API service lets developers engineer computer programs that interact directly with the AdWords server. With the applications created, advertisers and third parties can more efficiently - and creatively - manage their large AdWords accounts and campaigns.

From the Google Blog:

The AdWords API beta program is an open invitation to developers to explore new concepts (and then write great software) for managing Google AdWords advertising campaigns. Large advertisers can use it for their complex ad management needs, like tying product margins to optimized keyword bids.Third parties can use the API to build new interfaces to manage their client accounts. Best of all, an API enables the creation of all sorts of unanticipated ideas.

Radiation detection portals

Some of you might have seen the CBP (Customs and Border Patrol) announcement today re: the deployment of radiation detection portals at borders. The idea is to interdict trafficking of nuclear materials (among others) across US borders. If these are effective, border patrol have the ability to intercept nuclear weapons as they are brought across the borders. This would obviously be a good thing.

How might such a detection system work? I'm going to discuss this in very general terms because I have some misgivings on revealing potentially sensitive information. Consider a uranium bomb with, say, 12 kg of weapons-grade uranium and tungsten "tamper" that acts as a radiation shield. In a sense, this is a conservative weapons model (derived from Fetter, et al) - it is more likely that a terrorist group would use a gun-type bomb which would require about 50kg or more of highly-enriched uranium.

Such a bomb would emit neutrons and gamma rays, but the number of emissions observable at a detector may be smaller than the background rate of neutrons/gamma rays coming from cosmic rays, natural radioactivity, etc. So this presents an interesting problem of resolving signal from noise.

How can you make this detection problem easier? One obvious way is to move the detectors closer to the sources. Another is to increase the exposure time. To explain the latter point, consider a source that generates 20 neutrons/sec at the detector. The neutron background is 50/second with a standard deviation of 7/second (assuming a Poisson process with standard deviation equal to half the mean). Now if you see counts per second of 70, 75, 68, 75, 70..., you might notice a trend of 2-sigma events and conclude that there is a neutron source in your field emitting about 20 neutrons/sec. Well, the same goes for 1-sigma events, over a larger number of intervals, since the probabilities are multiplicative: a string of counts such as 59, 61, 64, 60, 59, 63, 56, 58, 60,... for instance, might lead you to conclude that what you're seeing is a smaller but still definite number of counts (perhaps 8-9 neutrons/second) above the background. So, given longer exposure times, it is possible to definitively detect weaker sources of radiation.

The truck or vehicle pulls up to or passes through the portal (a few meters wide) at pedestrian speeds (say 5 mph). This provides proximity and exposure time, aiding detection. Even so, this is a tricky problem, as noted earlier. Further, maximizing detection time is at odds with the goal of increasing throughput by reducing delays.

The above remarks primarily apply to passive detection, which consists of passively measuring gamma/neutron counts and registering counts that exceed a specified threshold. There is also active detection which involves actively probing the contents of a truck or car using gamma rays or x-rays and using the results to infer the presence of nuclear materials (this is conceptually similar to taking an x-ray image). This works quite a bit better, but obviously, since this is an invasive procedure that could affect any humans within the vehicle, this technique is not as popular as passive detection. However, this technique might be feasible at border checkpoints, where it might be feasible to require the passengers to step out of the vehicle for the duration of the inspection.