Out-of-band data center management via power bus

1. A method, comprising:

receiving, with a given rack computing unit, direct current (DC) power via a DC bus connected to a plurality of rack computing units of a rack, wherein the DC bus is configured to deliver DC power to the plurality of rack computing units;

determining that a duration of time designated for the given rack computing unit to access the DC power bus for transmission is occurring;

in response to the determination, applying a time-varying electrical stimulus to the DC power bus, wherein the time-varying electrical stimulus encodes an address on the DC power bus of a rack controller and a sensor measurement indicative of operation of the given rack computing unit;

concurrent with applying the time-varying electrical stimulus, filtering voltage fluctuations of the DC power bus resulting from the time-varying electrical stimulus to produce electrical power used by the given rack computing unit, wherein the filtering is implemented with a low-pass filter that outputs the electrical power used by the given rack computing unit;

determining that another rack computing unit on the DC power bus is transmitting concurrent with applying the time-varying electrical stimulus or another time-varying electrical stimulus and, in response, selecting a different duration of time based on a random value;

encoding with pulse-width modulation the address on the DC power bus of the rack controller and the sensor measurement indicative of operation of the given rack computing unit in the time-varying electrical stimulus applied to the DC bus; and

controlling the plurality of rack computing units of the rack with the rack controller via multi-point connectivity on the DC bus by encoding RS-232 commands with pulse-width modulation of time-varying electrical stimulus applied to the DC bus, wherein the DC bus is distinct from an in-band network connected to the plurality of rack computing units.

2. The method of claim 1 , wherein the time-varying electrical stimulus comprises varying a resistance of the DC bus.

3. The method of claim 1 , wherein the time-varying electrical stimulus comprises varying an impedance of the bus.

4. The method of claim 1 , wherein the time-varying electrical stimulus comprises varying a current drawn from the bus or driven onto the bus.

5. The method of claim 1 , wherein the DC power bus includes a generally straight conductive member extending along a side of the rack and having an exposed electrically conductive surface that delivers power to a contact on each rack computing unit.

6. The method of claim 1 , comprising establishing a remote serial console session via the DC power bus to manage the given rack computing unit.

7. The method of claim 1 , comprising encoding a signal indicative of the physical location of the given rack computing unit in the rack in the time-varying electrical stimulus or another time-varying electrical stimulus applied to the DC bus.

8. The method of claim 1 , comprising:

executing an application in respective instances of an operating system on the plurality of the rack computing units.

9. The method of claim 1 , comprising:

communicating with the given rack computing unit via a first network that conveys data the given rack computing unit operates upon; and

communicating with the given rack computing unit via a second network by which the given rack computing unit is managed, wherein:

the first network is different from the second network, and

the second network includes the DC bus.

10. The method of claim 1 , wherein:

the low-pass filter comprises an inductor in series between the DC bus and the given rack computing unit.

11. The method of claim 1 , wherein:

the low-pass filter comprises a resistor in series between the DC bus and the given rack computing unit; and

the low-pass filter comprises a capacitor in parallel with the resistor.

12. The method of claim 1 , wherein:

the low-pass filter comprises means for filtering.

13. The method of claim 1 , comprising:

receiving a data terminal ready signal by comparing the pulse-width modulated encoding to an unmodulated signal of a same underlying frequency of the pulse-width modulated encoding; and

based on the comparing, detecting variations in edge timing of the pulse-width modulated encoding indicative of binary values.

14. The method of claim 1 , comprising communicating, via the DC bus, signals encoded with pulse-width modulation, the signals comprising:

a data terminal ready signal, indicating that data terminal equipment is ready to receive, initiate, or continue communication via the DC bus;

a data set ready signal, indicating that data circuit-terminating equipment is ready to receive commands or data;

a request to send signal, indicating that data terminal equipment requests data circuit-terminating equipment to prepare to transmit data;

a request to receive signal, indicating that data terminal equipment is ready to receive data from data circuit-terminating equipment;

a transmitted data signal, carrying data from data terminal equipment to data circuit-terminating equipment; and

a received data signal, carrying data from data circuit-terminating equipment to data terminal equipment.

15. The method of claim 1 , wherein:

controlling comprises updating or querying firmware versions.

16. The method of claim 1 , wherein:

the address is part of an address space of the rack on the DC bus;

the address space is a different address space from that of DC busses on other racks.

17. The method of claim 1 , comprising:

steps for encoding pulse-width modulation.