Actually, there is a physical mechanism in cellular life that sets a lifespan limit, by roughly constraining the number of replications most differentiated cell lines in a body can support. Thus the correlation between slower metabolism and longer lifespan. The 'number of heartbeats' observation is a consequence of that correlation.
The mechanism is a little too complicated to describe accurately here. You can find a great deal on it with google search terms "telomere fraying" or "telomere fuse".
Very roughly, every chromosome has a 'telomere' structure at each end that acts to prevent the DNA double helix strands from 'fraying'. In the single cell embryo these structures are several thousand bases long. In every cell division after that, some random number of the telomere bases fail to be replicated. So the telomeres get shorter and shorter with each replication. In a cell where the telomere is all gone on a chromosome, genes at the end of that chromosome get progressively destroyed - thus (depending whether it's an active gene) wrecking some aspect of the cell's metabolism. With many different chromosomes (eg 26 in humans), 2 ends per chromosome, shifting gene arrangements, huge statistical populations of cells with individual telomere failures, and multiple lines of differentiated cells all replicating at differing rates, it's no wonder 'aging' shows such complex and variable symptoms.
I suspect if someone did a study of multiple species examining embryonic telomere lengths, average telomere loss per cell replication, replication (metabolic) rate, and average lifespan, there'd be a function of those factors giving a much more precise connection than 'number of heartbeats'.
On the first chart, we see that humans are definitely an outlier. Slide us back to the left and see where we would "naturally" be without modern medicine and we should live around 35 years. This lines up nicely with historical data[1].
Life expectancy at birth of both humans and Neanderthals pre-agriculture (Upper Paleolithic), is also just over 30 years. That works out to be almost exactly 1 billion heartbeats and correlates well with the regressions, according to those charts.
Given the above, it is kind of amazing how long humans remain fertile, even without modern medicine [worrying sigh - parents just don't tell you these things].
My point is that life expectancy is a useless metric in this context as here we're interested in average life span of a person that survived childhood.
In July 1969, just before the first manned lunar landing, Life magazine had a special issue about the astronauts. Many of the astronauts were profiled. All of them did some sort of athletics --- running, racketball, etc. --- except for Neil Armstrong, who was quoted as saying (paraphrasing from memory): I believe every man comes into this world with a finite number of heartbeats allocated to him, and I'm damned if I'm going to waste any of mine running around doing exercises.
Then Life magazine was among the misattributors; I read the quote when it was published (and was delighted, because it matched my own attitude about athletics).
That "stat" surprised me as well -- small dogs @ 10 years and large ones at 17? Uh, no.
Great danes and Irish wolfhounds (and other big breeds as well, AFAIK) only average something sad like 7 years, whereas those little yappy toy poodles average twice that.
Along those lines, smaller versions of the same species live longer than large versions. Dogs as you point out, but also cats, and humans. Look at mortality rates for 7 footers as a good example.
I feel for the hamsters. I'd have a 400+ bpm too if I knew I'd only live for 3 years.
But being serious, I'd like to see the the variance on those numbers. Bpm's of ~60 are normal for some, while others hover around ~90+. So within the range of what's considered a normal heartrate (60 - 100), numbers can vary by as much as 66%.
Some notes: Birds follow a slightly different power law and consequently live much longer. However, they do still follow a similar relationship.
Interesting exceptions in mammals are humans and (surprisingly) bats, both of which beat the curve pretty dramatically. Several mechanisms have been proposed for this exception including brain size and several specific proteins.
On a pet peeve tangent... I dislike the "Six Sigma" meme because, having a semimechanical heart, one-in-a-million failures would mean I'd be dead by now. I need one-in-a-billion reliability at minimum.
This was a 'fun fact' I was told at a young age. I'm not ashamed to admit that I took it literally and spent a few months of my childhood fearing anything that made my heart race.
Cardiovascular exercise leads to a lower resting heart rate, which more than makes up for the short term increase in heart rate during the exercise. Elite runners often have very low resting heart rates, like in the 20s-30s. So you should run, a lot.
This would be a fun one to work through. If 100 hours of working out reduced your heart rate by 1bpm (I have no idea if this is true, just sounds plausible), but during workouts you increased your BPM by 20%, where is the equilibrium?
I think you will need to refine your model. At the least, you will have to have some way to limit resting heart rate to be > 0. Alternatively, your model suggests that person will die from 6000 hours of practice, as it would bring his resting heart rate down to zero. If you don't, your optimization run would say something like "work out for 1,000,000 hours at a cost of 100,000,000 heart beats or so. Then, enjoy an infinite life with negative resting heart rate and even lower heart rate during exercise."
You also will want to model the fact that your resting heart rate does not stay at a lower level forever after exercise.
Finally, but that's peanuts compared to the other problems: let's start with a starting rest rate of 60 BPM. Exercise for 1000 hours to bring it back to 50 BPM, and your heart rate during workouts becomes 1.2 x 50 BPM = 60 BPM? Unlikely.
While resting heart rate is determined mostly by how fit you are, and is usually in the 50-80 range, maximal heart rate is generally a factor of your age. The accepted formula for maximal heart rate is (210 - age). Most workouts are at 60-75% of maximal heart rate (depending on workout intensity), so a 20% increase during workout is not likely (or at least, you wouldn't call it a workout).
Edit: fixed a typo.