Visit Your Local PBS Station PBS Home PBS Home Programs A-Z TV Schedules Watch Video Donate Shop PBS Search PBS
I, Cringely - The Survival of the Nerdiest with Robert X. Cringely
Search I,Cringely:

The Pulpit
The Pulpit

<< [ Shooting Ourselves in the Foot ]   |  Voter Flawed  |   [ Son of Napster ] >>

Weekly Column

Voter Flawed: A Follow-Up on Several Back Columns and Can Diebold Voting Machines Really Be Hacked?

Status: [CLOSED]
By Robert X. Cringely

Though some of my friends thought last week's column about security and other problems with the FBI's CALEA digital wiretapping system could have landed me in a Federal holding cell, that didn't happen. Nor did I expect it. The Feds have too much to worry about from their own poorly designed systems, leaving little time for bozos like me. The problem is that these systems are built in reaction to external events, and often the problem they are supposed to react to doesn't even still exist by the time the systems are built. Even sadder, the government procurement process is such that some equipment is obsolete before it is even delivered. What outfit is America's biggest user of Pentium-II computers? Uncle Sam.

Next week, I'll be writing a column that will make me enemies at the Recording Industry Association of America as I reveal a loophole that could allow some company to quite legally destroy the record industry while simultaneously making a ton of money. THIS IS REALLY IMPORTANT. But for this week, I need to follow up on some issues raised over the last couple of weeks.

Last week's CALEA column did result in readers sending other interesting/appalling stories, including one about Carnivore, the FBI's system for tapping ISP communications with you and me. The DCS-1000 or Carnivore system is apparently rife with security defects, starting with the fact that it is a Windows 2000 box exposed to the Internet, typically not behind the firewall, at the ISP and remote-controlled from the FBI office using PCAnywhere. The data it captures are downloaded insecurely in the PCAnywhere session. In fact, the FBI admitted that some significant e-mail intercepts concerning Osama Bin Laden were "contaminated" and were not legally usable (the technician reportedly was ordered to destroy all the intercepts) due to technical problems with the box.

The National Security Agency reportedly argued for putting a secure version of Linux in the Carnivore boxes, but that recommendation encountered mysterious administrative roadblocks. The fact that Microsoft was chosen just this week by the Department of Homeland Security to provide $90-million in software for the agency's 140,000 desktops and servers does not make me feel any safer.

And speaking of Microsoft, a story in the Inquirer this week pointed out that Redmond's office in the UK uses Linux to power its web site, which is supported by Akamai servers running the hated (by Microsoft) OS. These are the same folks, remember, who announced the ill-fated iLoo Internet toilet. But if you look inside the story there is something very interesting, which is the fact that Microsoft essentially did this to itself and will have a very hard time separating itself from Linux for edge caching, where Linux is king and Windows is hardly even a factor.

The big story here is that Microsoft messed up big time by supporting Akamai, but not SUPPORTING Akamai, driving the company, instead, into a sweetheart deal with IBM. Microsoft squeezes its vendors, demanding deals that often allow no profit while IBM leaves a little breathing room. Ultimately, the IBM deal probably involves many times the revenue of the Akamai's previous deal involving Microsoft software. Now Akamai is running IBM's Websphere on Linux, and if Microsoft wants a snappy web site, they are pretty much stuck with it.

Another interesting story appeared this week in my inbox from New Zealand claiming that Diebold voting machines in the U.S. (Diebold apparently makes most of the voting machines used in the U.S.) have major security flaws that allow manipulation of elections. These flaws are not so much hackable as they are designed into the system for deliberate manipulation of election results, claim the authors. I have no idea whether this claim is true or not, though the authors provided vast amounts of supporting evidence including source code. What is interesting to me is not so much that this could happen, but that we haven't read about it in the mainstream press. I didn't even bother investigating the story because it was sent to every reporter the authors could find. I figured that before I could verify anything the story would be in the Washington Post, yet it isn't. It isn't anywhere other than on a few obscure web pages and right here. It seemed to me to be newsworthy even if all the Post and the New York Times and the other big boys simply chose to debunk the story, yet they haven't done that.

Hopefully, the story is false. You can learn more about it under the Links of the Week on this page. If it is true, then it may well be the case that massive voter fraud has put many of the wrong candidates in office, meaning we aren't a nation of laws at all. Even more disturbing is the fact that the mainstream press doesn't appear to be interested, which is scary. You be the judge.

And finally, my story two weeks ago about TelePulse and its DTMD technology for high-speed Internet over phone lines, and the column two weeks earlier about Pulse-Link and its Ultra Wide Band overlay network for cable TV systems, brought some very interesting comments from Scott Kozicki, founder of Bluestar Communications, an early DSL company he sold to Covad.

According to Scott, who has long fought the fight, the telecom monopoly doesn't want any of this to happen. Despite the economic carnage of flat rate pricing, it is in their best interest to keep charging people for what they already have and do nothing more. The depreciation rates drop their investment to zero and profits rise over time. This is the formula that telecom has used for nearly 100 years. This little blip of investor turmoil is nothing compared to 100 years of monopoly, and just gives them a chance to fortify their positions and consolidate. More consolidation just means more of the same here. And we all know that real innovation happens with the small guys.

Kozicki argued that Pulse-Link's plan to work with cable companies is flawed because it is not in the best interest of the cable provider to expand their service offerings. They would like to continue charging us $50 per month for their current (and not so good) broadband service simply because there are no good alternatives today. There is just not enough incentive to give more value to the customer.

A better alternative for Pulse-Link would be to go around the cable TV company completely, claims Kozicki. Work with the consumer electronics people to build a device for the end user to use on the cable plant. (Imagine the cable plant as a transmission medium to talk directly to another user -- a peer-to-peer network instead of a broadcast network, effectively bypassing the provider.) But that is apparently not part of Pulse-Link's plan. There are already laws in several states prohibiting you from connecting your own devices to the cable system. Remember how we used to not be allowed to own our own telephones? It is the same kind of deal. The companies are already lining up against these types of technologies unless they are deployed totally in their control.

In Koziki's view, DTMD has much more potential IF IT WORKS, but TelePulse seems to be pursuing the wrong sort of partners. First, they've got to get the local phone companies to approve its deployment. Those companies will likely test DTMD for eons before they will give approval. This will give them time to fill the market with FUD, just like ISDN, and just like xDSL, until they get their product and marketing plans underway. The FCC needs to force this down their throats. Having Michael Powell sit on the board of the Telecommunications Development Fund, which invested in TelePulse, might help. But given Powell's ineptitude in nurturing a competitive landscape, it might not.

If DTMD gets approved, the Bells will attempt to either embrace it or kill it, and likely both at the same time. The Bells can play games at this level for years if they need to. They could decide that they want DTMD for themselves, so they'll make the tariff structure, deployment structure, colocation agreements, provisioning and all the other bureaucratic processes around procuring loop for a competitor impossible to navigate (like they did for xDSL). If they decide to kill it, they could simply put more fiber in the loop (FITL) and not allow competitors access to remote fiber fed terminals (like they did for xDSL). Even if they want to deploy it for themselves, the roll out rate will be glacial. The financial reality for the Bells is that they are under crushing debt, with no vision for the future on how to achieve a stable, high margin revenue stream. They only introduce a new product if forced to. Amazing how hard it is to be a monopoly these days.

Even if it all works out and DTMD is available as xDSL is for existing competitors and monopolies alike, the rollout might still be abysmally slow. Investors are still burned from the (still smoldering) telecom crash, and still don't have a fundamental understanding of what is necessary to create a large service provider network that makes money. They will likely be hesitant to invest in new carriers, and existing carriers will not be able to fund upgrades. Even if the money does come, new carriers and existing carriers will likely make the same mistake that they did this last go-around. They'll focus on the pipes and not the water.

Wall Street's myopic view of telecom players just fuels all this mess because they keep driving plant-based metrics instead of value-based metrics for carriers. It's not the number of lines or central offices or cities you have wired up that ultimately makes the difference. It's what people use those lines for and how you can derive revenue from those services. Cable has a better understanding of this, but for some strange reason, this thinking doesn't translate when you start talking about telecom. If telecom companies *really* offered advanced services like VoD, VoIP, VPN, storage, unified messaging, managed security, etc., they'd have a much stickier relationship with the customer while also deriving more revenue for each line sold. Investors don't really understand this still.

So the best approach for TelePulse is probably not to work with the ILECs, but with real facility-based CLECs. These competitive carriers have access to the equipment racks down at the phone company and can probably install DTMD. These CLECs need new products and new technology in order to survive against the big monopolies, so they have the only clear incentive to push something like DTMD. And if they do, the ILECs will eventually do it, too, just as they did with DSL. But it is going to take time, not because the technology isn't doable, but because there will be so much foot dragging. Sad to say, we'll likely have the same old story for at least another decade.

Comments from the Tribe

Status: [CLOSED] read all comments (0)