Over the last couple of years, applications of blockchain technology outside of cryptocurrencies were hard to come by, despite the prevalence of the buzzword fueling its rise up the Gartner Hype Cycle. However, after thorough exploration by startups and established firms like IBM, blockchains are slowly embedding themselves into practical integrations with major industries and governments.
In Mexico, where the government just recently partnered with a blockchain firm for grain tracking, the Federal District Government Transport and Highway Secretary of the state of Coahuila has announced its partnership with DeepCloud AI — a decentralized cloud computing platform blending blockchain technology and artificial intelligence (AI).
The goal? To bring tamper-proof, digital registration of vehicles to the nation with the MexiCar app, built on top of DeepCloud AI’s infrastructure.
“The MexiCar app is a ground-up implementation on a decentralized, low-latency cloud that solves many connectivity issues in rural areas and low-density areas,” detailed Max Rye, CEO of DeepCloud AI in the press release. “By adding the blockchain layer, we can ensure that third parties, including government employees, cannot change or manipulate the data.”
Decentralized Cloud Computing & Digitizing Government Services
Mexico’s recent forays into blockchain tech for supply chain management and digital vehicle registration are indicative of a favorable approach to the applications of blockchain technology. In particular, the digitization of government services that rely on the security and privacy assurances of the underlying blockchain for protection against fraud and data manipulation.
“Although it may seem this process is redundant to existing solutions, you can use it as a secondary source of verification as the transition to completely digitizing government services,” says Rye.
Trivial and transparent verification of data is one of the principal value propositions of blockchains. For example, in bitcoin, transactions are near-impossible to forge, but verification by full nodes remains simple, capable of automatic performance by a node with a canonical history of the blockchain.
However, the approach taken by DeepCloud AI directly targets cloud service marketplaces, more specifically, the IoT industry and edge device computation.
“Our AI-driven cloud uses AI and machine learning to optimize and match resource providers with app users,” says Rye. “Our cloud computing fabric shows better latency than anything on the market, and allows us to quickly deploy resources in regions that use the sharing economy to leverage network resources at the edge.”
At a high level, DeepCloud AI is a cloud computing network like Amazon’s AWS but, critically, with a distributed marketplace for resources like bandwidth, computation, or storage. As a result, costs and barriers to accessing cloud services are reduced, and the AI of the platform efficiently allocates those resources where needed.
“Our core differentiator is the use of AI for doing the resource matching between the network resource providers and application developers,” says Rye. “Because our cloud is up to 20x cheaper than AWS/Google, we can use the money saved to extend the network and additional services on the app.”
Application developers can subsequently build powerful, low latency applications at the edge — like with IoT devices and identifiers embedded in vehicles. Combined with a low-cost infrastructure and secure data technology using Intel’s SGX, and DeepCloud AI offers a compelling digitization of Mexico’s vehicle tracking.
“MexiCar builds on the tamper-proof capabilities of our blockchain infrastructure to secure vehicle registration via digital documentation with third parties such as government agencies, law enforcement, and insurance companies,” details Rye. “The tracking capability can deter potential crimes, mitigate corruption, and streamline reporting issues during accidents.”
Fraud and lack of accident reporting are endemic issues in vehicle registration, especially in areas where record-keeping and IT infrastructure are subpar. With MexiCar, law enforcement, government officials, and insurance companies can scan QR codes installed on license plates registered with the app for anything ranging from checkpoints to criminal investigations.
All of this is backed by the underlying assurance that the record is legitimate, but that is only the beginning of what DeepCloud AI and Mexico have in mind with the partnership.
Preparing for A Future of Autonomous, IoT-Laden Vehicles
The consistency of headlines focusing on autonomous vehicles from Tesla to Uber makes it appear inevitable that we are rapidly approaching the inflection point of autonomous cars on American roads.
Mexico’s move with DeepCloud AI is a forward-thinking push that may furnish the type of infrastructure necessary for autonomous vehicles to eventually thrive in the country. The signal of their projecting a future of autonomous vehicles is buried in the details of their partnership with DeepCloud AI though.
“Part of MexiCar is our strategic partnership with Xilinx, who is the leader in FPGA accelerator cards,” says DeepCloud CTO, Geeta Chauhan. “We are working with them to develop the world’s first-ever accelerated super-app — MexiCar.”
According to DeepCloudAI, Xilinx’s FPGA accelerator cards are 90x faster than CPUs, have 4x more inference throughput, and 3x better latency than GPU solutions.
“Clusters of FPGA cards will be distributed throughout the region to handle heavy compute workloads,” continues Chauhan. “Other apps do not have this technical advantage, which bodes well for supporting the requirements of autonomous vehicles and AI processing at the edge.”
DeepCloud has even taken the forward-thinking dynamic a step further, citing that they have “future-proofed” the technology underscoring the app. “Our research team has developed a proprietary, post-quantum encryption algorithm that will be integrated into the DeepWall security chip (i.e., similar to Intel SGX) embedded onto the Xilinx FGPA accelerator cards, which secures the DeepCloud fabric from the most intense attacks.”
A futuristic take, indeed.
As for the initial stages of the MexiCar rollout, the Mexican government and DeepCloud AI have some more proximal timeline projections in mind.
“We expect to launch a prototype of the vehicle-registration blockchain in the state of Coahuila by the end of July 2019,” says Rye. “The platform will integrate QR-embedded license plates on a nationwide scale by December 2019.”