No idea, but they need to withstand 2 volts reverse bias. Then they need to conduct with 2 volt forward bias from the pulse network. In about 350 ps.
Naively applying I=CdV/dT, you get a large number for dV/dT, around 11.5E9. So if you swing 2pF 4 volts in 350ps, you need about 20mA from the pulse network.
The 1N5711 is specified at 2pF junction capacitance, but that is at 0 volts. With 2 volts reverse bias, I am guessing from the graph it's more like 1pF. In any case, it's lower.
"But Snapdiode", I hear you say, "the 1N5711 is only rated for 15mA current!" But that's continuous, the 1S1 blasts this current pulse for probably much less than 350pS and at a quite low PRR.
Now Tek says the diode quad was "matched". In the days of hand-made diodes I assume this was to select for diodes with similar capacitance.
I'm going to theorize that modern manufacturing processes for the 1N5711 make them all very nearly identical.
But still, that 15mA number bugs me. Does anyone have 25 cents to spare to give this a try?