AMA recap: June 29 2018
AMA transcript with CEO Joseph Denne and the senior team, covering mainnet and Founding Node Programme
Question 1: When will I be able to host a website or store data on DADI network? **
Website hosting will be possible with release of DADI Web and storing data with the release of DADI Store. Check out roadmaps here: https://dadi.cloud/en/roadmap/. But you can use DADI CDN as a customer today! Reach out on firstname.lastname@example.org and we can discuss getting you set up… CDN features can be seen here: https://dadi.cloud/en/cdn/
Question 2: What are the specs of the Stargate in the photo you shared on Twitter?
Pretty punchy! 32x Xeon E3 1260L v5 Quad-Core @ 2.9GHz, 8Mb Cache / 1.2TB RAM / 2TB SSD raid. Plus, room to expand that 10x per node as requirements grow.
Question 3: You made the decision to not have a burn for transactions, what was the reason behind this?
We’ve never spoken about a burn for transactions. We originally explored the potential for a burn tied to DNS, however this would have been an artificial construct with no utility purpose in the network. As such it was abandoned (in early January, pre ICO).
Question 4: Will everyone be able to run a masternode if meets the minimum tokens requirement, or are there limited numbers and FCFS?
For the Beta phase numbers are limited. This is because while the network is early stage we see a need to balance the market (supply and demand), ensuring that Nodes are fairly rewarded. Without a level of control, we could easily find an oversupply vs. customer demand. Once we move out of Beta (there’s no firm date for this yet) we will remove controlled on-boarding and leave supply and demand up to the market.
Question 5: Constellation is just around the corner! Any prices on the packages being offered? What will be the cheapest? Who are the targets at that low level?
Pricing is per dApp. Pricing for DADI CDN — our first dApp — can be seen in the Masternode system document here: http://dadi.link/mC
Question 6: What can we expect from DADI Store? What are the use cases? How will data be stored? Will it be totally redundant across the network or only redundant on some nodes? How will it impact node requirement?
Store is a P2P decentralized filesystem, with files chunked and stored on trusted nodes within the network. Trusted nodes may include Gateways as well as Hosts, but certainly there will be some proportion of nodes in the network that will store the data continuously.
All other nodes that contribute to the storage layer will locate data on the network whenever they receive a request for a file/files that they don’t currently have themselves. After the data is served to the requester, that node will cache the data for a certain period of time (or evict less frequently requested data to ensure the node’s configured storage maximum is not breached).
Store could be used for a number of things, from storing holiday snapshots to distributing large datasets for analysis. We’re currently working on integrating it within the network as a decentralized database for API, a shared image cache for CDN and even the delivery mechanism for entire websites.
Question 6b: Do you think it will be quick enough to stream media?
When used in conjunction with DADI CDN, yes, we believe so.
Question 7: Why is the map on the new dadi.cloud homepage showing a demo of the network status rather than a ‘real’ data feed?
The testnet was a live hook up, and the new homepage map will be too. It’s going to a while to get it in place however, because we need to get the feed secured and also thinned out so as not to overwhelm the browser. Right now, the feed pushes everything in the network, which as you can imagine, even for a single site is a lot of data!
Question 8: A while ago I suggested including your partners on dadi.cloud– any news on this? Will those you are partnered with be willing to do the same?
Yes — Our update to dadi.cloud yesterday added partners to the community page: https://dadi.cloud/en/community/ And yes, our partners plan to do the same as well in time. We will also be adding detail on how we are working with our partners as each of the projects mature.
Question 9: How will you be onboarding nodes to the network? Will this be in a controlled manner?
Yes. Today we announced our Founding Node Programme, which represents the first phase of onboarding for nodes onto the network — selection will be by lottery and applications are now open.
This will be followed by further phases as outlined on our roadmap at dadi.cloud. We’ll be revealing more details on these in due course. Some useful links:
Founding Node Programme: https://dadi.cloud/en/updates/announcements/founding-node-programme/
Applications for the Founding Node lottery: https://dadi.cloud/en/contribute/
DADI roadmap: https://dadi.cloud/en/roadmap/
Question 10: Isn’t it a bit too complicated for a regular person to buy the tokens needed and to set up the node? On top of that there is an up-front investment needed and there is the problem with volatility as well. Why do you think an average person will go through all that only to make like 10$ per month from his laptop’s node for example?
Key thing for me is that individuals manage to set up Nest products, and to install software on their devices like Office or Spotify. It’s all about user experience and design. This is a huge issue in crypto at the moment: no one takes this side of things seriously and very few are doing anything to move the blockchain in to mainstream use. I’m happy to say that it’s a core focus for us, and always has been. Hopefully this shines through in our site, documentation and product design work.
We have a dedicated design team here, led by Dave Longworth, who is the ex-design lead for Monocle (monocle.com), widely recognised for its aesthetic.
With regard to proof of stake, it’s wrong to think of this as investment. It’s more akin to savings, with a return way and above what you can get with a bank. Of course, volatility is an issue, but we’ve been working hard to keep $DADI as stable as possible (given the wider market conditions), something that should be obvious from our price charts.
Question 11: As the ETH network is used for reputation management and accounting purposes we understand congestion is not a huge concern, but what point could it become a problem and what is the plan B?
The level of congestion required for this be an issue is huge — existential threat to Ethereum level. We’ll monitor performance, but do not expect this to be an issue. If it ever was however, of course we’d look to alternatives.
Question 12: 76 million tokens as circulating supply… wasn’t it 78M the other day? 2M staked?
In short, yes. Staking has been in testing, and we’ve deployed a number of nodes.
Question 13: How long has the Founding Node been in development?
Four months. The industrial design and fabrication process is pretty long, and even longer when dealing with custom chipsets. More details here: https://dadi.cloud/en/updates/network/founding-node-programme/ Oh, and can I say we have more products in development? Too late, I already did. :)
Question 14: Can you share your thoughts on how the DADI token economy has been designed? What factors will drive up demand for tokens in the mid-long term, bearing in mind that ‘increase in token velocity does NOT necessarily increase the value of network?
The key factors are decreasing token supply and increasing demand. Token supply is reduced as the network scales out, as POS takes tokens out of circulating supply. Token demand increases as customers are on-boarded, with a real time fiat on ramp driving purchases of $DADI from traditional businesses. And demand of course is driven by functionality and our sales support cycle, which kicks off in the coming weeks. Regarding devices, it’d like to see it in a smart fridge: DADI Cool™.
Question 15: With IEEE adopting the OpenFog consortium reference architecture for fog computing, plus other players in this space playing an active role in the development of these standards, where does DADI fit in? Does DADI’s approach align with what they’re trying to achieve? Could they be helping to shape possible standards or is this not worth the effort at this stage?
I personally see the OpenFog consortium as being virtuous, but premature. Why? Fog computing at this stage is in no way analogous to underlying technological protocols like TCP or UDP. It’s a nascent space that should be focused on rapid evolution and improvement, not standardisation. So, no, I don’t think it’s anything but a distraction at this stage — well meaning, but ultimately more of a method for old businesses to appear aligned to the new than an essential process to aid adoption.
Question 16: How big is the Founding Node? Will it fit in my house?
Hah! Here’s one of the prototypes in my hand: Gives you a good sense of scale.
Question 17: Who is the DADI client moving to the mainnet?
Big question. It’s Monocle. They are already on the network ready CDN and will be moving traffic over incrementally from next week.
Actually, this is a good opportunity to show you some of its features live. Eduardo has been putting together a post on it — perhaps use some images from Monocle’s site and drop in the examples we’ve been working on @eduardoboucas
What’s happening: making a square version (600x600) of a portrait image, using the entropy crop mode to keep the focal point in the resulting image
Just to be clear guys, the links above are to live images on Monocle.com running through network ready CDN. The URL paramters can be tweeked. Play with them a bit and you’ll get a feel for how it works. E.g. width down to 400px
Here’s a link to the DADI CDN Sandbox if you’d like to have a play with it: https://docs.dadi.cloud/sandbox/dadi-cdn
What’s happening: reducing the saturation to 0 to create a black&white effect + reducing the width to 900px
What’s happening: applying a duotone filter with a highlight and shadow colours + changing the width to 1000px
Layout plugin: https://cdn.somedomain.tech/layout/i:/butler_academy_netherlands_mon-5af3271b07133.jpg,h_600,w_300,x_0,y_0,l_0,t_0%7Ci:/180424_rhodes_stevenpinker_018-5afc44ae6f94a.jpg,h_600,w_300,x_300,y_0,l_0,t_0%7Co:5b2aad440c02e21eeca731b3.jpg,h_600,w_600
What’s happening: making a collage with two images, specifying in the URL their dimensions and positions
And for some interactive demos on Codepen (not using Monocle images):
- Lazy-loading a large image with a placeholder: https://codepen.io/eduardoboucas/full/RLOVGm/
What’s happening: a small blurry version of the header image is automatically generated by CDN and loaded onto the page. when the actual image, much larger in size, finishes loading (simulated in the demo with an artificial timeout), the placeholder is replaced by the large image.
- Lazy-loading an image gallery with placeholders: https://codepen.io/eduardoboucas/full/OxGVqb/
What’s happening: same as 1), except the technique is applied to a gallery of images and the placeholders are only replaced once the user scrolls past them.
- Automatic image art direction with theelement: https://codepen.io/eduardoboucas/full/ZOApOK/
What’s happening: aelement is used to switch between multiple variations of an image depending on the size of the viewport. these images are automatically generated by CDN, using the content-aware cropping tool, so that the focal point of the image is always shown on every aspect ratio, without requiring any additional effort from the developer.
Question 18: If DADI were the Beatles, who is Ringo?
Funny. How to answer without getting in trouble! Actually remember this one for a while back –
“In defence of Ringo Starr — a masterful drummer and the Beatles’ unsung genius”: https://www.theguardian.com/music/2017/sep/19/ringo-starr-great-drummer-the-beatles-genius
With that in mind, it would probably have to be Jo (Biddulph) — mostly behind the scenes here, but absolutely instrumental to the smooth operation of the business.
Question 19: Are there published performance results for DADI-hosted sites using CDN? To help us onboard new customers.
We have published some data that compares the request throughput between CDN 2.0 and 3.0 (https://dadi.cloud/en/updates/web-services/introducing-cdn-3.0#performance-improvements). In addition to this, we’ll be sharing various performance metrics as the mainnet evolves.
Question 20: What would be the minimum internet bandwidth requirements for a node?
That partly depends on location. Whilst some of us are lucky enough to have tens, hundreds or even thousands of MB/s lines, there are some still many less fortunate users. In those areas with less connectivity, where there’s a far greater distance from a well-connected area, it makes sense to allow Nodes and end users to connect locally, rather than forcing an end user, on a poor connection, to connect to a Host in another country.
Question 21: With the Founding Node lottery, how will we know the draw is random? Are you going to do it live via random generator?
Yes, exactly as we did for the whitelist at ICO. We’re publishing the source code of the randomiser as well.
Question 22: If I’m not selected in the lottery, can I buy the Founding Node? It’s a really cool design.
No, the Founding Node is a limited edition. But there will be other devices in future 🦄
Question 23: To ensure that nodes are geographically spread out, will there be preference to node applicants from lesser known locations?
Our global backbone aims to offer full coverage. We’ll be highly incentivising availability, and working closely with local data centres to make the best use of any redundancy they have.
Question 24: I believe it was mentioned that DADI tokens won’t be the only form of payment. Other crypto will be accepted too? Why will people choose to use DADI tokens when there are other tokens that are more widely used?
That’s correct — we will take payment in multiple currencies. Payment in another currency will trigger a real time conversion to DADI, so there’s no reduction in demand for DADI as a result. We are also implementing a direct fiat on/off ramp with a new partner that will allow instant conversion to DADI.
Question 25: Out of interest, how many nodes would you estimate will be online by the end of next year?
Impossible to answer as this will be driven by customer demand in the main. However, I expect it to be a lot. A metric lot.
Question 26: Any timeline for when Agorai deep learning nodes will run on the DADI network?
We have a co-dev timeline in place for their use of the marketplace which Jim (our VP of Technology) is leading. We’ll be able to share date in due course.
Question 27: If I am in an area that doesn’t have any nearby Hosts or Gateways, will the fallback be a traditional data centre?
This is very similar to question 23, but we have contingency in the network via our backbone so will never need to rely on old school cloud services to bail us out. In time I expect they’ll learn to rely on us. :)
I always find it funny when “decentralized’ platforms end up being run in existing cloud infrastructure. Hardly decentralized is it?
Question 28: I made some calculations based on Digital Ocean droplets that reflect roughly the needed hardware. And I translated all of it to $ according to the current DADI Token price. This shows a loss for a Gateway and a Host overall. Will it ever be possible to make profit with a Gateway or a Host?
Benchmarking in a VPC misses the point somewhat, although you can run in them no problem, and if you pick the right provider and spec, will turn a profit.
Note that the figures quoted are minimum. At capacity Stargates and Gateways will return several times the quoted percentages.
Question 29: I know things are still being tweaked, but how much sysadmin will be required for a gateway? e.g. Will we be running Docker instances and managing a Linux distro ourselves?
The network applications are automatically updated, can handle scheduled downtime and recover after network or power outage. In terms of sysadmin, the only required work would be machine updates and bare metal upgrades. You can request an update window in order to avoid penalties for unexpected downtime.
Next AMA is on July 13.