Reading Into Microsoft’s Acquisition of Fungible

Reading Into Microsoft’s Acquisition of Fungible to Accelerate Datacenter Innovation

Microsoft announces acquisition of Fungible to accelerate datacenter innovation

“Today’s announcement further signals Microsoft’s commitment to long-term differentiated investments in our datacenter infrastructure, which enhances our broad range of technologies and offerings including offloading, improving latency, increasing datacenter server density, optimizing energy efficiency and reducing costs.”

Discussion:

You have to kind of read in between the lines on this article for two reasons, both of which hint at Microsoft’s possible future planning:

1.??????The announcement uses the term “DPU” (data processing unit).

2.??????They also mention “disaggregated” architecture.

Data Processing Unit (DPU)

One of the selling points that DPU (data processing unit) offers is that it optimizes performance for use over networks.?Also, one interpretation of DPU is that it’s a more generic term for all forms of “processing units”, with the more common ones being:

• CPU – (central processing unit) used in datacenters for general data processing using the “traditional” central processing unit (CPU).
• GPU – the slightly newer graphics processing unit (GPU) used for parallel data processing.

But…

• QPU – The quantum processing unit (QPU) is up-and-coming and will appear in many different forms. And after QPU, who knows what new xPU will come next?

“Disaggregate” Component Advantages

One of the selling points that Fungible’s DPU (data processing unit) offers is its “disaggregate” (non-centralized) component architecture.?In other words, components (such as processors) and data storage (like SSDs) don’t need to be physically located near each other anymore.

According to Fungible’s website, this lends to a “fast fabric” architecture that allows anything to be connected without compromising speed. The following depicts processors and data storage media working independently of each other.

Formerly, all processing unit components were built into one unit, and were optimized to work only with one another.?Now multiple disparate components can work together.

However, you can also read a few other possible advantages into this:

• Resource load balancing – By decentralizing (disaggregating) these components, it bears obvious back-end advantages like load balancing and resource redundancy (in the event of data disaster / data loss).
• “Swappability” – As the components are physically separated from each other, when comes time to swap out one or more of them, their non-interdependent form factors might allow for easier swappability. This would reduce swap-out downtime, reduced operational costs, and easily facilitate upgrades.
• Modularization – This modularization allows for the above-mentioned swappability (mentioned above) not only for upgrades, but also for “mission specific” or “purpose built” tasks. The “mission specific” modularity aspect would be especially appealing to military clients.

For example, the US Navy builds “mission specific” submarines. Some are pure weapons platforms. Some have few to no weapons, but carry cargo instead. These submarines have a modular hull section that can be swapped out according to the mission at-hand. One modular section that remains the same: the propulsion section, as all subs need to move.

In similar fashion, the “disaggregated” nature of the Fast Fabric architecture allows for easily modularization.

• Transition to Quantum – The “Fast Fabric” graphic above physically emulates one of the main tenets of the newest science: quantum entanglement. In the quantum sciences, “entanglement” means literally anything can be connected. Kind of like in IoT, but much broader as to include everything. (This entanglement idea even shows up in movies now!).

The reason that “disaggregate” components are transitional to quantum is that it allows for easy swapping into new tech like quantum. In similar fashion, a swappable, modular “fast fabric” architecture would allow for entanglement-like capabilities – a kind of technology that may serve to help customers bridge resources between “classical” computing resources (like CPU and GPU) in a hybrid architecture connecting to a quantum system (using QPU).?

***

John Melendez is a technical writer who began his writing career 15+ years ago as an editor and worked upward to author tech content for MICROSOFT, GOOGLE (Taiwan), INTEL, HITACHI, and YAHOO! He now advances his knowledge through self-study in the Quantum Sciences and Space Tech. He speaks intermediate Mandarin Chinese (會講國語) after living for 10 years in Taiwan, Singapore and China.

John Melendez has qualified as an FAA Airman (single-engine land), FCC Technician Class HAM Radio Operator, Lt. Col. (ret.) CAF, SCUBA, Snow Avalanche Rescue Cert.

#quantum #quantumcomputing #space #aerospace #electrogravitics #QaaS #QCaaS #technicalwriting #techwriting #writing #writer #editor #chineselanguage #mandarin #taiwan