TEA Project Competitors

Comparison to Dfinity

The main difference between the TEA project and Dfinity is that Dfinity does not rely on secure hardware. It instead uses cryptography and consensus to ensure its subnets are tamper-proof, which has both advantages and disadvantages. Dfinity can be mathematically proven to be secure but still consumes excess computing resources. The amount of computational overhead needed means that it’s not suitable for edge nodes to participate. Dfinity also relies on the existing cloud computing infrastructure to distribute its load. General public miners are not likely to participate, and potential risks are re-centralized.

The TEA Project instead relies on cheap security chips and does not require expensive CPU, GPU, ASIC, hard disk drives, or other expensive equipment. TEA mining machines are currently positioned as single-board computers with a TPM security chip and a GPS module. The TEA mining machine solution ends up costing less than $100 as it doesn’t require high-energy complex cryptographic computing and consensus algorithms. This makes it convenient for the public to participate in mining. In the future, TEA machines may even be built into set-top box routers or even refrigerators. This setup also avoids the risks of re-centralization.

The TEA Project relies on hardware (TPM) to perform the decryption step whereas Dfinity relies on software. This creates a need for server-class hardware with low latency to mine on the Dfinity network. If however only datacenters are suitable to be mining nodes on the network, then that introduces a new type of centralization risk to the platform. The TEA Project’s mining nodes can be purchased for a modest amount, probably less than $150. This allows for a large, diversified mining infrastructure that’s topologically more distributed than clustered datacenter nodes.

Dfinity is similar to the TEA Project in that both make use of Wasm and its inherent flexibility as far as being language-agnostic. But whereas the Tea Project uses the same 3-tier architecture as cloud computing, Dfinity does require more investment on the developer’s part in learning their underlying tech stack and how things are done differently in the Dfinity world.

• Although Dfinity canisters can communicate with each other as long as the code’s compile target is Wasm, the Motoko language is heavily favored by current devs on their platform.
• Motoko was designed with Dfinity’s concept of orthogonal persistence in mind. Over time, the expectation would be that developers for Internet Computer do not use MySQL databases or even files as they learn to take advantage of memory persistence instead.

We at the TEA Project want developers to get started writing dApps without having to study the underlying architecture or learn a new coding methodology. Using the common 3-tier architecture means that more developers can onboard seamlessly onto our platform as an entryway to Web3. More developers = more chances of a killer Web3 dApp being built on our platform!

Comparison to IPFS / FileCoin

IPFS is only responsible for providing storage and network services, while FIL needs to consume hard drives to compete for mining. TEA Project runs on IPFS, utilizing the storage and network provided by IPFS to extend trust computing features for IPFS.

In fact, TEA is an ideal add-on module of IPFS. It can rely on the established IPFS miners to quickly form a computing power market. IPFS miners can add cheap TEA mining machines to dual mine on both the TEA network and IPFS.

IPFS and the TEA Project are not competitors. The TEA Project relies on IPFS and provides computational power for IPFS. This will allow IPFS to become a truly decentralized Internet computing node rather than just storage.

Comparison to Other CPU Trusted Hardware Projects (iEXEC, Phala, etc)

The blockchain projects using a trusted hardware layer generally all use CPU technology such as Intel’s SGX. These projects include iEXEC and Phala. These are some of the main differences that the TEA Project has with these projects:

1. CPU-reliant projects are dependent upon Intel to provide the security certification. This makes Intel part of the central security authority, a solution that is not thoroughly decentralized enough. TEA Project does not use CPU technology; instead, it uses extremely cheap TPM chips, a Trusted Platform Module. The TPM technology has been around since 2003, giving it nearly 20 years to streamline its design and decentralize its distribution through the many companies that manufacture the chip.
2. The Trusted Execution Environment (TEE) programming for the CPU is only accessible on its bottom layer. The average application developer cannot master such bottom layer technology. The security architecture of Intel CPUs thus hinders the participation of ordinary application developers. The TEA Project does not require developers to understand any of the underlying hardware. It only requires them to use common programming languages to compile their program functions into the Web Assembly format and reprint them on TEA nodes for calculations. This hardware-agnostic design of the TEA Project greatly reduces any barriers to entry.
3. Intel’s CPU cost and power consumption are both very high. Generally speaking, all computers produced after 2008 have TPM chips. Because of its simple design and availability, the cost to run a computer with a TPM chip is meager. These machines are flexible enough to be used in miniature low-power IoT devices. This simplicity of design unlocks real nationwide edge computing that can be done on devices with TPMs.
4. The TPM chip used by the TEA Project does not directly participate in computations. Therefore, many specialized heterogeneous computing types can be used. For example, the specialized TensorFlow processor (TPU) for artificial intelligence, the image processor GPU, and the new Risc-V processor can be used on TPM-protected hardware, but Intel CPUs cannot coexist with these specialized processors.

Relying on Intel for the security certification places too much power in the hands of Intel, a form of centralization that carries risks. Indeed, Intel recently announced it would be deprecating SGX in the 11th and 12th generation of its consumer CPUs.