Greetings, colleagues! In today’s article we analyze the super-KANT in details: what is it, what is it for and how it functions. Have a productive reading!
What is super-KANT?
SUPER-KANT is a unique link (module), connecting all KANT modules from individual “islands of the economy” to a single balanced GONT economy.
As an example, imagine a motherboard of your computer. The motherboard is the backbone of the computer. It has various connectors for cards, where you can install video, audio, modem, network card, etc. Thus, the motherboard connects all components into one. Exactly the same as super-KANT!
The basis of SUPER-KANT is the transport system GOYA , which is based on the unique VAVILON (ASIP-based) cores. In fact, now the transport system GOYA 2.0 is being written.
The reasons for the “Kant disengagement” in GONT may be different. For example, the owner of very cheap electricity in a geographic region (a monopoly on a cheap resource) will always strive to own his KANT module, in order to make services in a completely different economy (in their regional domains). But to participate in transactions in the global economy through SUPER-KANT is also very profitable for him.
SUPER-KANT provides a service (on a physical, transport level) for “leveling” and mixing KANT-economies.
SUPER-KANT provides an additional strong innovation.
For example, a full “hardware protection of consensus” will be implemented only in Super-KANT.
By the time the transition to the chip implementation is ready, several independent KANT subsystems will already be working in GONT and, possibly, there will be several variants of economies (according to quite different emission models).
At the same time, any new service can choose the best for itself (service) KANT module. The mechanism of KANT-competition will work.
Therefore, the time will come for global unification.
Strong innovation is needed to prevent “wild forks” GONT, which will become economically unprofitable. It is more advantageous to unite according to the KANT-compliance principle.
Therefore, SUPER-KANT should be:
- Strongly innovative (new types of cores and tires).
- Strongly protected (new cryptography technologists).
- Strongly adaptive and scalable (vehicle acceleration system, special instructions).
SUPER-KANT should be much more complicated than the possibilities of usual development teams to make fork GONT and start dividing the network into independent clusters (into sets of noninteracting economies).
SUPER-KANT should become the fundamental basis for combining all types of gVM cores on the market.
It is possible to implement only on the chip (HW) level.
In this case, the specification for chips will continue to be written on the basis of studying the work of conventional KANT (which can take 1-2-3 years).
If we consider services on GONT as an analogue of the usual branches of the economy (as we now believe for the formulation of the Cantonomy equations), then SUPER-KANT will strengthen the exchange processes between the GONT branches. A greater degree of transactionality in SUPER-KANT will lead to even greater earnings.
Motivation for SUPER-KANT:
The main technological value of GONT is a protected (including hardware) distributed container of virtual machines GVM, the continuous integrity of which is the fundamental basis for the long-term development and existence of GONT. Including, the guarantee of rapid expansion of the container GVM (and this builds the entire system GONT-ICO).
The SUPER-KANT transport system should realize all the variety of different variants of Value Chain within one service.
Levels of trust in channels in the GVM container (g-Channels).
Aim: The value of the chains (based on g-Channels) in GONT has a trust level no lower than trust chains in Ripple. In this case, GONT will be attractive for investments.
Models of transport
SUPER-KANT should be based on the latest models of transport! And such examples of models of SUPER-Transport, like Wave (+ MIPS) are already “almost” on the market.
Wave (+ MIPS) is a system based on CGRA. Ideally similar to FPGA, but the data go in groups and much faster. Tens of thousands of processor-like elements on the chip.
This is very similar to the concept of GONT.
The era of the implementation of SUPER-KANT.
Dialectics of coming to SUPER-KANT.
SUPER-KANT should lead GONT to the final point of evolution (in its current sense) – to the concept of “Datacenter as a Computer” and “Hosting 2.0”.
When the blockchain system will change or will begin to strongly influence the architecture of data centers and the ways of selling hosting services.
In this paradigm, the GONT must move on in a few years.
From this point of view, investments in GONT are short term trend.
Investments in GONT are investments in long-term and stable growth.
Protective mechanisms of SUPER-KANT
The GVM container becomes an independent product unit and can be connected through secure interfaces to various BCs.
The need to store secret keys (protecting the Web3 communication channel for the “Blockchain-GVM” bundle) leads to the need for super-protected sgVM cores. It is needed for the cryptographic service of this channel (including, for the implementation of the protected key storage) in the GVM on the side of the miner (which can be a “wild miner” and begin to crack the protection).
And only on the HW (chip) level implementation of sgVM protection will become absolutely reliable.
On the side of the same GONT-Ethereum client, we can only work with public keys (in the framework of elliptical cryptography).
Implementation on the model GETH AND PARITY. GONT’s own client – GONTE
Model: maintaining full compatibility with the current state of Ethereum (full-fledged Ethereum client) + the possibility of Cantonomy. This is the GONT client.
The whole logic of working with the virtual machine is separated from the main code Ethereum. The virtual machine has become a GVM container.
But, nevertheless, we need to support the work of Ethereum client as if nothing has changed for external observers. Therefore, it is necessary to enter your own client GONTE (thereum).
Transport system of SUPER-KANT
Dialectics of the transport system GOYA:
From service processors  to parallel start of virtual transport channels [Bill Dally], i.e. channels of added value (GONT Value Chains).
Why does SUPER-KANT need it and what does it protect?
First of all, virtual channels (g-Channels), built dynamically for the passage of transactions through gVM cores [ref-Bill Dally]. Here we compete with the channels in Ripple.
SUPER-KANT also has a Tile-architecture
The extensive use of Tile-architectures in the industry began 10 years ago and emerged from significant scientific research.
An example is Tilera.
In GONT Tile, this architecture is a “backing” for building gVM interaction channels. Channels continuously “reset” states for either intermediate storage or for storage in the BC.
In many respects, the Tile architecture has a duty to work with intermediate transaction states.
gTile also encapsulates the operation of the core interaction protocol (GPS protocol).
Hardware consensus protection is also implemented through gTile.
In the process of building a channel, there is a need to speed up many operations. For example, the operation of verifying the trust between the cores (hash like algorithm) can be accelerated in gTile.
Features of TILE:
- Specificity of transport to application domains (ASTRO system) can be realized through gTILE.
- TILE architecture can be a hierarchical.
GONT requirements for HyperTile
- Fast exchange bus between the cores.
- Multiple accelerators of operations.
- Confirmation of integrity through CHESS (Bugis system).
- Accelerator Instructions for gTiles.
Instructions Accelerators can be built directly into the processor pipeline. GTile core has a certain set of instructions – accelerators. Specialized instructions for both processing traffic, and for cryptography and integrity checks.
Examples of accelerators:
In GONT, all objects (all gVMs) receive GOW addresses on the GONT-Tree tree (in the planes in which they are issued).
In the process of “advancement” of the service on gVM and gTile substrates from the cores (see below), intensive inter-service data exchange and corresponding routing on GOW addresses of the cores is possible.
Therefore (inevitably) a lot of TCAM accelerators and packet parsers will be used. And under this special instructions will be sharpened.
The need for intensive routing by GOW cores.
Let’s enter GOW-GONT Ontology word (Actually – Ontology code on GONT Tree).
For the initial hit in the GONT economy, you need to buy one or more GOW addresses. GOW address is the most basic GONT asset. At the same time, the ACC – GOW (Ethereum account of the investor on the address in GONT) is displayed.
Possession of GOW can be open and closed.
Conclusion (by accelerators):
Through the system GASM (GONT address space management) in FUPE GONT space TCAM blocks are entered.
The gVM cores are addressed by pairs <Account; GOW Addr>
The owner’s account appears on the GOW address of the core.
In the process of transaction processing, intensive routing (g-Routing) takes place, which needs to be accelerated, including through TCAM.
g-Routing and DC-Routing (routing of the data center level)
When going to the stage 5 of the development of the transport system (to the concept of “Datacenter as a Computer” and “Hosting 2.0”), there will be a need not only to optimize local routing between gVM cores, but also global routing in the data center.
Having the global goal of GONT to realize “Hosting 2.0” and become more efficient economy in comparison with usual hosting (current hosting in Amazon and Yandex), we also offer the architecture of new routers for the GONT transport system.
Differences between Tile and gVM
- Tiles can be quite small – 1000.
- gVM can be unlimited – 1 million or more.
Tiles are real, permanently-connected kernels for implementing data transport.
gTiles “make” a fast transport for GVM.
The GTiles array is also introduced via KANT gas (SUPER-KANT gas).
Thank you for attention! See you soon.