Hi guys. Just took this message from Radix Network telegram.

There, they compare their network to competitors including Constellation. This is what they think of DAG. Read patiently,


“#Constellation
@Mvatansever78
I continued with researching them and though I was banned on their trader channel I got some more insights.
Their network Hypergraph is a little bit different than other DLTs, it can be best decribed as validating the data of p2p end nodes sent to each other (kind of data streams).
This includes their token which is kind of a demo data structure sent between nodes.
The messages / data is stored on the network (in I think a scalable way). Therefore, this means Hypergraph "L_0" in itself works like a layer 1 dlt.
The state channels / micro services they talk about are more like nodes / endpoints / bots than being part of the main network.
Especially, these state channels are run by "domain experts" (=controlled by someone). What the network does is controlling/validating the messages sent from and to state channels, but not the state channels themself.
Lattice btw also runs in one state channel (probably controlled by the Lattice team).
Also nobody could confirm whether it is possible to do transactions with multiple state channels in one atomic transaction.
Their main claim is "being the layer 0 for all layer 1", what they mean is that they can connect all of these and act like a bridge.
However, they have a very insightful video about the cross-chain / L0 stuff: https://www.youtube.com/watch?v=0T7zcv_ieTc
Basically, how it seems to work is that (e.g.) 3 bitcoin nodes are also running a Hypergraph node and wrapping it into Hypergraph (called by them "L_0"), then the swap would happen on L_0 and then bridged back to e.g. Ethereum (via 3 ethereum nodes).
Overall, seems like a centralised bridge (what prevents the 3 nodes from stealing the fund in their BTC/ETH wallets).
Especially, there was no talk about how to decentralise the bridge (or they consider this as decentralised) like RenVM with darknodes. Compared to RenVM there was not much.

Finality seems quite high (because transactions can be propagated through the network for confirmation) - but don't have numbers.

Overall, it is an interesting concept (for the data streams), but has some weak spots like their probably centralised bridges and that state channels are not really secured by the network (only the messages sent from/to).

I hope this helps, they make a lot of bold claims though like smart contracts are centralised and dead in a few years. The main problem I have with this statement is that one part of smart contracts is that they are immutable (or most are / can be) and it seems that the code/behaviour of these state channels could be modified at runtime.

Take this with a grain of salt, their architecture is a bit different and I got banned questioning their "L_0"^^, since I said calling it "L_0" is a bit like a marketing phrase because in itself it is like other layer 1 DLTs (though they want to connect others, but on the other hand then I also could call other scalable DLTs layer 0 because they have no problems connecting other dlts with centralised bridges). What is new though that you can interact with bot-like wallets (state channels).
Anyway, I will follow up with them later.”

I am hoping to receive comments from DAG chads who has in depth knowledge to the technology.

Hope someone could give feedback on this analysis

I agree. It’s good to see the perspective of other people, better from competitors themselves.

Lot to unpack here but basically they don't fully comprehend what a state channel is, the topology of the network, or how data enters the network. Basically all of these notions and inferences will be resolved when we update all of our documentation and whitepaper.I can understand they tried their best to do a competitive analysis with the information they have. The hypergraph is a DAG network that connects independent DAG networks together (referred to as state channels). The hypergraph itself reflects the global state while each state channel reflects a local states. The local states have custom validation/consensus criteria and finality. These local states converge into the global state with its own, typically longer finality window of 20 seconds or so. Nodes are performing the actual validation. As data enters the network through state channels, nodes are validating the data itself. Everything is being oraclized and normalized into a data structure that can be understood by all other state channels, meaning there is atomic composability throughout the network. There are multiple consensus algorithms running concurrently that validate both the data on the network as well as nodes themselves through proof of reputable observation. State channels are NOT centralized bridges. They are essentially layer 1 DAGs with programmable logic and are a native component of the multi-tiered topology of the network and inherit the trust of the layer 0 Hypergraph. State channels can handle any complex data type and are not relegated to only those that traditional smart contracts support which inherit their limitations from their respective layer 1 networks data type support. To put it simply, Constellation is validating data in real time while layer 1 networks are validating transactions occuring on a ledger. Constellation is using Java/Scala, a functional programming language, coupled with a hylomorphic recursion scheme. State channels are validated at compile time, not just run time. The atomic composability at the Layer 0 is feeless, enabling the Layer 1 state channels to not be constrained or impacted by any price actions occurring with $DAG, the layer 0 asset. Basically state channels have economic independence while still being natively connected to the greater network of state channels. On the question of legitimacy regarding the USAF contract: They are a paying customer. Constellation is basically working in Phase 2 and Phase 3 in parallel. Getting tech integrated with the US DOD is a methodical and sensitive process that requires adapting to security requirements and the nuances of the network environment. Everything is progressing as far the Constellation team can disclose within the pre-production staging environment (this is referred to as a pilot). Proof of concept was phase 1 and they are passed that now. Also to note, it's not just one agency involved from what has been communicated. It's considered a strategic multi-agency effort.

I cover everything here as well: https://link.medium.com/RaUVkzWBolb

Thank you I’ll definitely read this.

This is a good visual I made as well: https://twitter.com/HSethv/status/1465491620921491458?t=s_IUzzB_lshfHeRf2kuvOA&s=19

In haven’t seen this yet. Finally some visuals for better understanding! Thank you!

permission to send this? They are all curious about the DAG tech. Is the article only for dag chads?