Dynamic Bridge Tolls

On my way into work today, I was thinking about bridge tolls. There has been talk of raising the tolls on various Bay Area bridges to $4 or $5 or even $6! Yikes!

Now that we have FastTrack, it seems that there may be more creative solutions we can come up with. For instance, what if the toll is variable? Those crossing at 5am get to cross for $0.50. At 6am, its $1.5, at 7am it jumps to $3, at 8am its $6, at 9am it drops to $3, etc etc?

I found a similar proposal for the New York area from the Tri-State Transportation Campaign which had some good data in it here. They divided their plan into “off peak” and “on peak”, at $4 and $6 respectively.

Using excel, a little math, and a bit of arrogance, I am coming up with this model. My goals are to increase revenue, but also increase vehicles-per-hour over the bridge.

1) If you don’t use fasttrack its $7.50. This is because you are costing us a lot of money and slowing everyone down. Get fasttrack.

2) Peak hours (7am to 9am): $7

3) Off-Peak hours: $2

Now this is a no-comprimise plan with teeth! A commuter can save $25 per week by shifting to off-peak driving hours.

Using the same percentages for vehicles per hour as the TSTC used in New York, my model would produce a 5.7% increase in revenues over the current, static $3 plan. At the same time, it puts such a heavy incentive to get drivers onto FastTrack and to not use peak hours, I think we’d see a significant decrease in overall bridge wait times.

But what do I know….

Really Extreme

But what if we wanted to get really radical, and have a completely variable price that adjusts in real time?

Google is an interesting example of a company that adjusts their prices in real time. They deal with a huge volume of “ads”, and they are constantly figuring out which ad is best optimized for their business. This is done in real time.

What if all California bridges were linked electronically. As tolls are collected (either digitally via FastTrack or via toll-takers), the central system adds up the money. At the same time, the central system knows how many vehicles per hour are flowing through any bridge.

Each year, the governor types in a number for how much revenue per day he wants to get from the Bridge system. Then, the bridge system uses its historical models, current traffic flow rates, and the target revenues to calculate how much it should be charging drivers to cross the bridge at that instant.

Imagine the case where a bridge is extremely congested. The bridge system would automatically start raising the toll – in REAL TIME. This would have the effect of also lowering the tolls on the other non-trafficked bridges also in REAL TIME. We could set limits on the toll so that it never exceeded say, $15. But when it gets congested, commuters need to find another way to get home.

The advantage of this system would be that bridge tolls would be almost free when crossing if there was no traffic. But, when congestion occurs, the price would get higher and higher. Its actually a very fair way to deal with pricing.

Anyway, this idea is wild and crazy. The state of California isn’t very capable for creating real-time, dynamic bridge crossing toll algorithms. And we can get “close enough” with something simpler and less politically damaging anyway…

Operational Uptime

We’ve all heard companies that claim service reliability of “99.9%” or “99.99%” or “5 nines” (meaning 99.999%). But what does that mean?

The fact is that 99.999% uptime is really hard to achieve. On a monthly basis, 99.999% uptime means that your service is “down” for less than 26 seconds. Now, many services skate around this issue with fuzzy definitions of “down”. They might say that “downtime is defined as a majority of mailboxes being unavailable”. So a single mailbox being down might not be considered “downtime”. But as customers, if its your mailbox, that doesn’t sound quite right, does it? We usually have simpler definitions.

Anyway, GMail this morning has been down for 3 hours (at least) on my mailbox. Assuming they operate perfectly for the remainder of the month, that means they’ve dropped to only 99.58% availability this month. And, if they remain up with no downtime for the rest of the year, they’ve already missed the 5-nines, and are struggling at 99.96%! Given the outages I’ve seen, I think they’ve been down for at least 12 hours on my mailbox, so I think they are operating under the 99% level.

Oh well, I shouldn’t beat up on them too much. It is only beta, after all, right? But geez- with the $4B in the bank from their second IPO, I’d think they could ad least get to “two nines”. What about my god-given right to read email now?

Microsoft vs Google in Court

Lawyers spar at Microsoft-Google hearing

You know, this is such a waste of time. How many millions do we have to waste on this stuff? This is a lose-lose situation for Microsoft and Google. Both sides should figure out how to settle this immediately. Either Google should say they just don’t need this guy (because they don’t), or Microsoft should they don’t need this guy (because they don’t) and that they don’t care if he goes to Google (because it doesn’t matter).

That’s probably the best solution – both sides agree to never hire him again.

But the lawyers are advising both sides to take this to court. They probably advise that this is somehow precident setting for future employee hires. Yeah right. The real reason is because that’s how lawyers make money.

GMail down again

I’m really starting to question my use of gmail. It seems pretty unreliable – I can’t tell you how many times it reports, “Gmail is temporarily unavailable. Cross your fingers and try again in a few minutes. We’re sorry for the inconvenience.”

The humor was funny once, when this was an isolated event. But I want to read my email now, and I really don’t have time for this incompetence. Its been down for at least 30 mins now [6:30 to 7am]. Yahoo! Please take me back!

Update – its now 10am, and the service is still down. 3 hours of downtime on a Weds morning. Not good!