They are designed quite differently.
The high voltage power supply within flash units recharges a big capacitor slowly (a fraction of a second, or two). The SB-800 to about 325 volts, and the Alienbees to near 500 volts at full power. When triggered, the flash tube is connected directly across this capacitor, and the flash tube absorbs all the of capacitor energy nearly instantaneously, ionizing the xenon gas, which becomes bright and fast.
Studio flash units are slower because they always fully dump their capacitor when fired (and this takes a little too long for milk drops). When set to lower power settings, the capacitor is partially charged to a lower voltage, but it is always fully dumped. Because, the flash tube is connected directly across the capacitor, and when the flash tube is triggered to conduct current, it is a direct short across the capacitor, totally discharging it. The flash tube becomes very bright very quickly, and then trails off slowly (relatively).
This type of studio lighting is slower when used at the low power settings (which is the opposite of speedlight units). This can be seen, but it is not a large factor, maybe double longer. It is much more noticeable that low power units made with smaller capacitors are faster than high power units made with larger capacitors.
The AlienBees B800 is rated 320 watt-seconds, and the B400 is rated 160 watt-seconds. Flash watt-seconds (joules) of energy = 1/2 CV², where C is capacitor size (farads) and V is volts. The B800 uses double the capacitance (two capacitors) to double the power. When triggered, the flash tube light output is quickly full bright, and then it decays to zero (relatively slowly, like a millisecond) via the standard RC exponential curve until empty. The math of the standard RC exponential curve depends only on RC and time (search Google for: RC time constant). The R of the capacitor and flash tube may increase with less voltage and current, but basically what we can plan on is: if double Capacitance for double power, then a double discharge time. Or, if less voltage is used for less power, then a little slower discharge time (maybe twice longer at 1/32 power).
The AlienBees flash duration specifications say the B800 is rated 1/3200 second and the B400 is rated 1/6400 second, both measured using the t.5 method (which is a standard method, not without exception, but in the ISO and DIN specs, to assign one time value to this slow decay). That sounds fast, and it is required to say it that way because the standards define it, and the competitors also do it. At least AB says t.5 and explains what it means. But the t.5 is defined as the time duration that the flash output is above 50% power (or if we can ignore any insignificant rapid rise time, it is essentially the time to fall to 50% of the peak power). So this t.5 time duration only reduces the light one stop from maximum peak output, and you can see there is lots of decay left to continue blurring the image of the falling milk drop (half of its power), until it finally dies away weakly. It is convenient to measure t.5, but it is not possible to exactly specify the speed of any flash unit, since there is much opinion involved about when the capacitor discharge actually completes. For practical purposes, when is it finished? At 10%? At 5%? At 2%? We will wait a very long time for 1%. For this, engineers picked 50% as the standard number which defines a power response curve. But a better (more realistically accurate) practical time specification for photographers is t.1, to 1/10 output, which is numerically roughly three times longer time than t.5 (on the standard RC curve). It is a safe assumption for any RC curve that the t.1 duration will be about three times longer than the t.5 duration.
EDIT: AlienBees are also publishing the t.1 specification now. The number to compare to other flash brands is still necessarily the t.5 standard, but t.1 is the number that affects your picture. For example, Alienbees B400 full power t.5 spec is 1/6000 second, and its t.1 spec is 1/2000 second. The photos on the previous page clearly show its photographic actual is around halfway between the Nikon SB-800 1/2 power 1/1100 second and the 1/4 power 1/2700 second ratings (i.e., the B400 t.1 1/2000 second seems very correct photographically). This discussion is only about when we quit measuring output, at 50% or at 10%, it does not change what the light does.
The Excel graph shows the standard RC curve (Capacitor discharge time through a Resistance), which is well studied in electronics because it describes the way any capacitor will discharge. (This curve is output = e to the power of (-time/RC), where e is the "natural" constant approximately equal to 2.718.)
So for our purpose here, this curve shows how the power at any flash tube decreases relatively slowly, which affects the flash duration. The output of the triggered flash tube becomes immediately full bright, and then decays as shown by the blue line, from 1 (100%) to 0 output, along time units left to right - the entire graph could be 1/1000 second overall width for our flash tubes. The arbitrary t.5 and t.1 points are marked, indicating the time to decrease to 50% power and 10% power. It should be dim at t.1, but it takes a long time to go to zero. This curve shows the slow decay of the flash power, which explains the visible milk drop motion, which gradually fades away (shown on previous page). Half of the light still remains after it decays to this 50% point, and the rate of decrease slows too.
There are really not many options to design the studio light faster. The design consists only of a flash tube wired across a capacitor... two components. What would you change? Flash watt-seconds (joules) of energy = 1/2 CV². We could design with higher voltage and a smaller capacitor, or vice versa, but this changes the Xenon gas ionization spectrum (color temperature). We might find better components from better vendors, certainly all components are not equal. But the point is, all the rest of the flash circuitry is only about recharging and regulating the capacitor voltage, or triggering the tube to conduct. Which is very important, but which does not make it faster when it fires. The RC curve is electronic basics. The B400 has one capacitor, and the B800 has two, or 2C for 2x more power and 2x longer duration (shown clearly on previous page). The B1600 has four capacitors, for 4x power and 4x duration. Less C is faster, but is also less power. Less R is faster and the better components may have less internal resistance, but they are what they are. More expensive high power units (White Lightning X1600 is one) may switch in multiple capacitors, using more of them for high power, but using only one for low power, so it won't be so slow at low power (will be a very fast small flash, instead of a big slow flash turned down even slower). That is a big deal. However even then, they still do a full RC discharge, and are still too slow for milk drops.
The advantages of this full-dump direct flash tube connection are: no switch circuitry in the high power output (low R), and also a more consistent color balance (however all flash tubes vary color spectrum with different power levels through the flash tube). Studio lights are high power, and plenty fast for the studio lighting, but this is not fast enough for milk drops up close.
Conversely, camera speedlight units are much faster duration (at their lower powers), because they use either a thyristor or IGBT semiconductor as a switch, which simply opens the output circuit (disconnects the wire to the flash tube), abruptly terminating the flash tube current when the specified power level is reached. For example, the 1/4 power setting performs more like the green line marked on the graph above ... the full output begins, but is simply interrupted there, after only a very short time (1/4 power is when the Area under the curve is 1/4 of the total area). The capacitor is always fully charged, then triggered, and then interrupted. The lower the power, the faster it is... perhaps only 1/20,000 second duration, perhaps less. There is no remaining additional exponential decay, and the capacitor retains the unused charge so the next recycle is faster too. If you are into high speed photography, then this speed is a very big deal.
Here are excellent pictures of actual measurements for a Canon 430EZ speedlite, which is characteristic of all speedlights. This picture clearly shows the difference in the way that speedlights output their power. This also makes a dramatic difference in flash speed. It is why speedlight units are much faster at lower powers. This is why they are named speedlights. The disadvantage of course is the low power output, but which does not matter in many cases.
The Nikon SB-800 flash duration specifications are shown below (manual power levels, from Specifications chart, page 122). Note this is the time until the abrupt cutoff, with zero remaining output (except for full power of course).
At full power, the speedlight is not terminated early, so then it has the same full exponential decay as the studio lights (actually, even slower, the thyristor increases R at full power). Nikon does not specify if the full power speed rating is t.5, but t.5 is the standard industry-wide rating method for flash duration - it is always assumed. If it is t.5, then it ought to be nearly the same number as its sharp cutoff 1/2 power rating, even if it is an entirely different situation and result. The previous page photos showing the B800 and B400 speeds show us what a theoretical 2 to 1 difference looks like. In the same way, the speedlight at full power might be 2x slower than the B800 speed. However the speedlight enters an entirely different class at 1/4 power.
Nikon calls these Speedlights (SB-800, SB-600), and Canon calls them Speedlites (580EX, 430EX), and this is the reason for the name. Older models too, and most others of this type (used on a camera, instead of being studio lights) do the same thing, Vivitar or Metz for example. Not all models offer 1/128 power, but any should go down to 1/16 power. Technically, the Nikon SB-800 and SB-600 are not thyristor, they now use an insulated-gate bipolar transistor, IGBT, same switch concept but more modern. The lower power of small speedlights make it feasible to interrupt the output. So the speedlight can be extremely fast at low power, but the color balance will shift more between full power and lower power.
The speedlight becomes more blue at lower power. Any flash tube pulse starts out hot and blue, and trails off to cool and red as the voltage decays. Xenon flash tubes have the property that this color averages out over the full pulse duration to be the white that we call daylight. But cut off the trailing decay (for lower power), and the remaining initial part is more blue as it becomes shorter and faster. In contrast, the usual type of studio lights (as described) use the full pulse, but at lower voltage operation for low power, they become slower and cooler, a bit more red (the opposite of speedlights in both respects, speed and color). The Alienbees Einstein lights can combine both offsetting methods, one method compensating the other, for more consistent color over the power range, but this is quite expensive at high power levels - add another 0 digit to the price. The color consistency is not a big problem to the photographer, because the session setup uses only one power level, and we learn to handle the color. The power level and color of different lights can vary, but color is never exact anyway.
It is difficult to compare power of different flash units from Guide Number specs, because the coverage angle of the reflector makes a huge difference of how that light is concentrated. The Guide Numbers (GN) have fooled people into thinking the Nikon SB-800 is somehow stronger than the Alien Bees B800, because the Nikon SB-800 has GN 184 (105mm zoom, feet, ISO 100) and the AlienBees B800 has GN of 172 (standard 7 inch reflector). But no way the little Nikon is more powerful, because at 105mm, the SB-800 only covers a small area specified to be 27x20 degrees (specifications chart page 121), and the AlienBees covers 80 degrees with the 7 inch reflector. If the subject is inside that tiny area, yes, the GN are near equal brightness there. If outside the covered area, well, you know. But the speedlight has many GN values, one for each zoom angle. At 24mm zoom (78x60 degrees), the SB-800 has GN 98, which is nearly half, which is nearly two stops less than the 80 degree Alienbees reflector. If the AB B800 uses the optional 11 inch 50 degree reflector, its GN becomes 320, approaching double (and GN 320 is 3.3 stops more light than GN 98, which is about 10 times more light in the covered area). The area covered by the light is all important - area requires power to fill it evenly. We can only compare numbers for equal covered areas. So the point is, Guide Number is all about one reflector and one coverage angle. If you change that, then the old GN is meaningless, and you must measure with a flash meter to see what you have now.
However, in practice, if these two lights are individually placed on the same stand and same umbrella at full power, and metered at the same distance, the Nikon SB-800 (24mm zoom) meters about 1 stop less power than an AlienBees B400, and two stops less than a B800 (7 inch reflector). So in this umbrella use, I'd claim the SB-800 was perhaps 75 watt seconds equivalent in a metered comparison. That, plus the SB-800 1400uf capacitor and 325V voltage values compute to be: 1/2 CV² = 75 watt seconds.
But umbrellas are not pertinent to milk drops, direct flash from a few inches is a key factor there (at lowest power). At its 1/128 power setting, my estimate is about 0.5 watt-seconds for the SB-800. But it still works fine up close - maybe 7 to 10 inches (about 18 to 25 cm), ISO 200 and f/16 - and it sure is fast. (Note: 0.75 foot x f/16 = GN 12 at ISO 200, which is GN 8.5 at ISO 100, which agrees with the GN chart for 1/128 power at 24 mm).
In summary, the studio light capacitor is partially charged depending on the power level setting. Then it is always fully dumped when fired, and this is relatively slow. The speedlight capacitor is always fully charged to maximum level, but it is abruptly interrupted to produce lower power settings, and this is often quite fast - however then its comparative power level is rather low. All units are fast for portraits, but only the speedlight at low power is fast enough for milk drops.
I love my AlienBees, and certainly I am NOT knocking them in any way. I am merely mentioning how studio lights are made. They are built for studio situations and they do that very well. They are consistent, and very versatile with light modifiers, for example they are vastly better in a softbox (bare bulb inside there) than the fresnel lens speedlight could ever imagine (instead just use bounce with a speedlight - an umbrella is very good). But studio lights and speedlights are simply constructed very differently, each with advantages and disadvantages. The speedlight at low power is much faster for high speed photography of milk drops.
Continued on next page, to the Setup to photograph the milk drops.