AMA Recap: Masternodes

🔗Question 1: Are stargate nodes load balanced clusters or can a stargate in a datacenter be a single server?

Adam K Dean:
Stargates are intended for high-connectivity environments: think data centers and high bandwidth office environments, and are designed to be single, powerful machines rather than a cluster of smaller, less powerful machines.

Arthur Mingard:
Worth noting that the Stargate application utilises all CPU cores, so in terms of machine performance it’s as efficient as possible.

🔗Question 2: In the deep dive it says ‘up to 20%’ for gateway and ‘5-10%’ for hosts. Payout is based on platform usage, right? So does this mean that hosts earn 5% minimum? And if so does this mean gateways could earn less than 5% if there is very little usage?

Arthur Mingard:
Gateways will never earn less than Hosts because there are multiple Hosts to a single Gateway. There could be anything up to 250 Hosts -possibly more - connected to a single Gateway, meaning that the Gateway will be taking a percentage from 250x more than a single Host.

🔗Question 3: If the node fails, is there a way to recover the staked tokens?

Joseph Denne:
Staked tokens are stored in the originating contract for the token, and the Nodes themselves are linked to an account within the network, accessed via dadi.cloud. This means that a failed host will never lead to a loss of tokens.

🔗Question 4: How do I know my network is fast enough?

Arthur Mingard:
The Host deployment process will run a number of availability benchmarking tests to give an accurate performance analysis prior to network on-boarding. This will indicate, amongst other things, the number of concurrent requests the network believes the Host can achieve, as well as an indication on payout, pegged to performance and demand, and with a view on suggested improvements.

Network performance will continue to be monitored on an ongoing basis to ensure that the DADI Network can adjust to fluctuations in network conditions.

🔗Question 5: Will programs on a Host be able to escape and get access to the local network/Internet, or does all the Host node traffic only go through a Gateway?

Adam K Dean:
Applications will run inside secure containers with restricted capabilities and with no access to an external network. Traffic between the Host service and the applications run over a private network. All of this will run inside a virtual machine running a hardened version of Alpine Linux (with Grsecurity).

The applications that are running inside these distributed containers will be DADI Webservices only, and will only allow inbound requests from the Host application. The Host application only makes requests that are queued in Gateway. All deployed builds of the DADI Webservices will be subject to security testing as part of the build process.

🔗Question 6: Staking and increasing customers will certainly drive up the price of tokens. Does this mean that DADI will begin charging more, in turn, in USD to reflect the increasing cost of DADI tokens?

Joseph Denne:
No. DADI’s services will be priced to remain competitive in market benchmarked in major currencies.

🔗Question 7: Can DADI Host really run on a Raspberry Pi?

Adam K Dean:
Yes. In fact the proof-of-concept network was hosted entirely on Model 2 and 3 RPi’s. If you take a look at the specs of the low to medium tier AWS, Azure and Google Cloud hardware, they’re very comparable, low powered machines.

🔗Question 8: When did you have the idea for the DADI Foundation?

Joseph Denne:
The Foundation was established in January this year, and has been in startup mode ever since. It has its own board and its own CEO. While it benefits from the network and was setup by the founders of DADI, it is independent, with its own articles and governance. You can find our more here: https://dadi.foundation/

Paul Regan:
And I should add that I’m with the CEO as I type attending the Blockchain for Social Impact conference in Washington DC.

🔗Question 9: Would websites for other crypto projects be a good target audience for early network adopters?

Joseph Denne:
We are in discussion with a number of projects and obviously love talking to anyone interested in our platform. We’re also looking at how we can directly engage the broader crypto community through the offering of free usage tiers for example.

Paul Regan:
If anyone reading is interested in talking about this, DM me.

🔗Follow up: This is great news. Are you able to comment are these large size crypto projects? Midsize? By marketcap?

Joseph Denne:
I can’t comment yet, but as soon as we can, we will.

🔗Question 10: Will there be a network/bandwidth/storage capacity measurement and monitoring tool released along with DADI staking details? Or would we rely on third party tools for these?

Adam K Dean:
We are working to build measurement and metrics in to our interfaces. In this first instance this will be via DADI CLI and also via dadi.cloud. In time we will be releasing graphical interfaces for MacOS, Windows and other platforms as well. Unfortunately, due to the security requirements, DADI services will be running inside of a secure virtual machine (a blackbox if you like), which will restrict you from performing your own measurements.

🔗Question 11: What does ‘peak usage’ mean?

Arthur Mingard:
In respect to the network peak usage, this is the peak potential throughput based on the
number of active hosted DADI webservice deployments and their respective request limits.

🔗Question 12: Any plans to make use of embedded/dedicated GPU power in the future?

Joseph Denne:
The use case for GPUs in DADI are tied predominately to our AI tools. The short answer then is
yes, but this is some time out.

🔗Question 13: Is there a mechanism to register early interest in running gateways in specific geographical locations?

Joseph Denne:
There will be, yes. We need a strong geographic spread, and will be on-boarding to aid this. Sign up on site to express your interest: https://dadi.cloud/en/join (And don’t forget to us which country and city you’re in!)

🔗Question 14: How do you monitor network performance. Do you have any local/global reference point for latency/packet drop?

Adam K Dean:
Stargates monitor the performance of Gateways, and Gateways monitor the performance of Hosts. This way, node performance is always representative of its actual performance within the network. Metrics such as latency, uplink speed, etc, are continuously gathered by the parent node.

🔗Question 15: Can I run DADI in a VPS?

Adam K Dean:
While running within VPS will not give you the same performance as running on bare metal, it will still be possible, yes.

🔗Question 16: When can I run a node?

Joseph Denne:
Our mainnet goes live next month. The first contributed nodes will be going live in July.

🔗Question 17: Payouts are monthly, so I wonder how they are calculated and paid out?

Arthur Mingard:
The share of the payout is calculated as a percentage of the total throughput, but also includes proof of availability weighting. So if one Host’s hardware is more valuable to the network than another, or the Host has a higher uptime, the payout will be balanced in their favour.

🔗Question 18: I would like to setup a Gateway. Can I put two servers in different places to have redundancy? Then if one is offline, the other one will work as a backup Gateway?

Adam K Dean:
No. Two machines cannot operate as a single network node. The network itself guarantees redundancy of applications through the distribution of application containers to multiple gateways and hosts.

🔗Question 19: In the tokenomics doc you say that the staking requirement will change depending on market demands. What do you mean by this?

Joseph Denne:
As consumer demand for the technology increases, the size of the network will need to increase. To support growth in capacity the requirement of POS will be reduced. The token value is the second factor that will be monitored and factored in to POS requirements.

🔗Question 20: You have a trusted by section on your site. Do you plan to have a partnered with section as well?

Joseph Denne:
Yes we will. And regarding exchanges, see: https://dadi.cloud/en/token

🔗Question 21: Is 20TB per month transfer fine for Gateway nodes?

Adam K Dean:
While utilising 100% of a 250 Mbit/s uplink would require 75.6 TB of traffic –

250 Mbit = 31.25 MB
31.25 60 = 1875 MB = 1.875 GB /minute 1.875 60 = 112.5 GB /hour
112.5 24 = 2700 GB = 2.7 TB /day
2.7 28 = 75.6 TB /month

In reality, it is unlikely that a Gateway node would utilise 100% of the available bandwidth 100% of the time. If the line was 25% utilised 100% of the time that would give you ~19 TB of traffic over a 28 day month. So, in summary, 20 TB should be fine, yes.

🔗Question 22: Dadi has no direct competitor at the moment. What are the plans for when competition arises from new projects or if big established companies enter as well? Are there any barriers to entry? Do you foresee any of the big companies as possible competitors in the near future?

Joseph Denne:
We’ve been working on our tech for just shy of five years, and the web services themselves have been in production at scale for most of these. So yes, there are barriers to entry. I think it will difficult for existing clouds to move to a decentralized setup as they are publicly listed companies. Moving to a democratised footing would fundamentally undermine their businesses proposition in market. The barrier for them isn’t about technology so much as it’s about business models.

Offering support for the blockchain is something that most are already doing, but this is a pretty cynical move. If you think about it, we have a bunch of decentralized projects in market today that are majority running on centralized services, returning profits for those highly traditional businesses.

DADI is the only viable network offering an alternative to this.

🔗Question 23: What are the criteria for selecting who can run a Stargate or a Gateway?

Arthur Mingard:
Onboarding will start with a FCFS application. Initially, we’ll be taking on Stargates and Gateways based on demand, and for Gateways this process will become automated. Hardware will undergo a number of availability benchmarking tests before whitelisting is complete.

🔗Question 24: How does the Fiat on-ramp work?

Joseph Denne:
Individuals and businesses can pay for DADI services using fiat if they want to.

This does not change the fact that our services are paid for in DADI. It simply means that there will be a small real-time exchange in place on dadi.cloud, removing the barrier to entry that a purely crypto based payments solution would pose. It’s no different in concept to a business heading to OKEx, buying tokens and then paying in $DADI, other than it is faster and that it provides the experience that the majority of our potential consumer base expects at this point in time.

Of course you will be able to buy our services in Bitcoin, Ethereum, Nano and many other currencies besides.

🔗Question 25: Are the apps made on the DADI Network anti-filtering? For example, if i create chat messenger on DADI tech can the government filter this?

Joseph Denne:
Decentralized apps are much harder for governments and other parties to filter, as they run from no set location. In this manner DADI works to help to support net neutrality.

It is important to state that we will be working with local jurisdictions to ensure that the platform remains compliant. Exactly what this looks like is a question for our legal team, and will shake out in the fullness of time. We can work with you to answer jurisdiction specific questions as they arise however. Reach out to an admin directly.

🔗Question 26: When participating as a Host, Gateway or Stargate you estimate revenue in % of stake from the POS. Is this the only revenue or do you also get compensated depending on how well your node serves the network? An awesome host machine doing heavy lifting must earn more than a weaker machine?

Arthur Mingard:
The participation payout mechanism takes network availability, connection latency and throughput limits, uptime and a number of other KPI’s so there is a considerably higher potential yield for a machine that appears to be technically greater in overall capacity. There are continuous tests, including peer review, to make sure each Host’s availability is consistent.

🔗Question 27: Can you give us an in-depth update on your marketing efforts, are they now well and truly underway?

Joseph Denne:
We’re spending pretty heavily on marketing and awareness. At the moment this is targeted largely outside of the crypto space, but this will start to shift in the coming weeks as we move to network on- boarding. I should note that network going live is the first real/legitimate marketing opportunity that we’ve had after the Crowdsale, as to promote to these communities before now would have been simply to promote token value (which is not something that we can do).

We’ve retained a blockchain marketing agency in recognition of the need to readdress this segment of our potential audience for example. We have a bunch of new videos in the works which will shake up our video content. Whether or not we maintain a channel is a good question that I’ll give some time to next week.

🔗Question 28: Is there any evidence DADI web services will be faster than others such as AWS or Azure?

Joseph Denne:
Plenty! You can benchmark for yourself by downloading the source of CDN today and taking it for a spin or by checking out some of the sites already on our tech.

🔗Question 29: Will you be able to switch Host hardware in the event of failure, or would you have to wait until the stake is returned after (up to) 30 days?

Adam K Dean:
Yes, this will be fine. Hosts have a minimum requirement of 20% uptime. If a Host goes offline, let’s say due to a hardware failure, once you’ve fixed it and turned it back on, it’ll reconnect to the network and resume its duties.