The VectorQ2 Power Controller provides Software-Defined Electricity for 3 phase, 208/240V power networks.
Microsoft data center goes down for several hours: https://www.geekwire.com/2018/microsoft-azures-southern-u-s-data-center-goes-hours-impacting-office365-active-directory-customers/
This is a fairly detailed article regarding the failure modes of a large data center in Wyoming. What would you describe as the top 3 advantages of VectorQ in mitigating failures of this type?
The three main advantages that Software-Defined Electricity will provide for this (and all) data centers are:
1. Awareness of the problem the instant it starts
Task Oriented Optimal Computing is the core data processing technology behind Software-Defined Electricity. It is what provides the efficient data processing capacity that is required to process data at speeds faster than electricity travels in order to control it in Real-Time.
This edge computing hardware technology that is installed into the power network is constantly building models of every load during normal operation. This is where the digital accuracy in data acquisition becomes critical. With 100% confidence in accuracy at the speeds and astonishingly high resolutions, the system builds a perfect power consumption baseline model for each load during normal operation. This is the precise performance signature at essentially the circuit board level which is always updated and maintained.
If this Microsoft data center were protected by Software-Defined Electricity, all of the loads would be individually modeled at this level prior to this event in the normal course of digital synchronization. This circuit board level modeling capability in Real-Time would allow the system to easily identify performance abnormalities that deviate from the precision load model.
This level of awareness at this moment in time allows for most problems to be predicted in advance. An example of this is the capacitors in HVACs or power supplies that simply wear out over time, consume more energy as they age and harm the power network when they fail. With Software-Defined Electricity, the data center operators would have known the capacitors were beginning to wear out months in advance.
Model based computation in the power network allows for any deviation from the model to be instantly identified. In the case of this Microsoft data center, we do not know the true specifics, but I am confident in suggesting that there were signs way before this occurred but they did not have the tools to detect them. Software-Defined Electricity is the only feasible way to accomplish this.
In fact, this methodology is also the ONLY WAY TO IMMEDIATELY IDENTIFY distributed malware that affects machines in a power network because the visibility required for attack identification must occur at the circuit board level and must be able to be instantly compared to a true baseline of the same. The same is true for deviation of processes.
2. In the moment optimization of power flow
The article refers to problems with the cooling equipment forcing a shutdown which drove the data center to preventatively shut down portions of the data center in an effort to prevent a much bigger disaster. This highlights the fragility of modern data center systems and how the HVAC is built to meet the precise capacity of the heat emitted from the power network.
Any time spent not cooling runs the risk of an out of control service outage. This is a crazy reality for a mission critical facility. When a facility is operating at capacity like this, even a simple event can have catastrophic consequences.
Software-Defined Electricity ensures that the power flow is always optimal, meaning that the current and voltage are always synchronized and perfectly balanced across the phases at all times no matter the load consumption or upstream fluctuation. This is critical for data center applications because it eliminates power quality as a variable and establishes it as a near perfect constant.
Always perfect power quality universally reduces energy consumption providing additional capacity to the power network and loads. In the long run, it opens opportunity for radical redesigns of future data centers because everything that is presently used in a modern data center has been overbuilt to exist in power networks with uncontrollable electrical fluctuation (larger generators, thicker wires, oversized components, etc.). Future data centers will be able to be much leaner in cooling and power infrastructure and more dense by IT assets.
Fluctuation and imbalance in power is frequently the root source of events that cascade out of control. It is responsible for falsely triggered events, brownouts, all the way up to full scale downtime and contributes heavily to the high levels of redundancy needed within data centers not to mention an enormous energy and resource waster.
The article does not mention the root cause of the HVAC failure, however if this Microsoft data center were protected with Software-Defined Electricity, no matter what the cause, the resulting event would be completely minimized, logged and the effects would not ripple through the network.
3. Eliminating power fluctuation during off grid transition
The most harmful and chaotic time in a data center’s operation is when power is interrupted and backup power procedures begin. Transitioning from grid power to any other source of power (i.e. flywheel, UPS, generator, battery, etc.) induces a lot of stress on the power network infrastructure and the loads. During the offgrid time, there is a significant shift in the heat profile of the data center, with an significant increase noticed in the power supplies, PDUs and UPSs.
What is occurring to cause this shift in heat profile, which is actually an increase in the energy wasted and a destabilizing force to the power network and loads, is the shift in power network impedance.
While connected to the grid, the impedance is low because there is virtually an infinite amount of power available, however immediately when the data center goes off grid, the impedance shoots up because the flywheel, UPS and/or generator does not have the same availability of power.
The result is the loads are essentially being starved of power. This is a significant reason that backup power is an absolute last resort, because it is inherently damaging to the operation in the short and long term.
If this Microsoft data center were protected with Software-Defined Electricity, it could be powered from any source whether grid, flywheel, UPS, generator, battery etc., without any fluctuation in power delivery before, during and after the transition. This occurs because part of the natural synchronization process is the automatic matching of impedance for the network which maintains the impedance close to zero at all times no matter the upstream or downstream fluctuation.
This is a significant contributor to downtime prevention because it prevents the harmful transitions and starving effects suffered during the off grid time. This allows the data center to operate without impact no matter the power source because perfect power quality and electricity availability are now one and the same.