AMA Recap: October 04th, 2019

Friday’s AMA was a broad range affair, covering subjects from legal compliance to the specifics of network architecture. Those questions not answered this time will be picked up in the next AMA.

Lead article image

🔗Question 1

🔗Will the dependencies that we use to install our nodes be replaced by decentralized solutions? I’ve read that Docker, for example, is backed by an In-Q-Tel venture which is publicly known to be a capital firm of the American CIA.?

Arthur Mingard: I can’t really comment on rumours around Docker and any CIA involvement, but we do have a general position on third-party dependencies - the fewer the better. We’ve already finished working on supporting Balena which we will introduce in the next quarter, and are also working on a binary solution that has no dependencies whatsoever!

🔗Question 2

🔗Why not give zip files to extract instead of new SD cards or port forwarding?

Arthur Mingard: Each device must be setup with a UUID so we can’t create a generic zip. The good news is we won’t have to update SDs again as they are now self-updating, just like self-onboarded devices.

🔗Question 3

🔗I’ve heard about voting rights. What are the specific things this will apply to and do you have an idea when and why the first voting will occur?

Joseph Denne: We’re going to be building a voting mechanism in to the heart of the network, which will give voting rights to all Stargates, and to the top performing Gateways and Hosts. (One vote one node.)
Voting rights will cover everything that has a material impact on the network, from key junctures in development to revenue distribution.

At the moment these sorts of decisions are taken by the team at Edge Network Technologies Ltd., the not for profit entity charged with the ongoing delivery the network. In time they will be broken out and voted on by participants in the network. This may also include customers. We like the idea of a customer board/panel being able to be part of this. There’s no timeframe in place yet. As soon as there, we’ll let you know.

🔗Question 4

🔗Will we be able to receive some kind of notification if a node, founding or self-onboarded, stops working?

Arthur Mingard: We are working on a solution to provide a number of notification channels, including email alerts, API subscriptions, Slack notifications and web hooks. We’ll be announcing plans around this shortly.

🔗Question 5

🔗Can we add a staked indicator to the explorer, for example 3.5% staked at the moment?

Joseph Denne: I love this. Yes. We’ll add it to the backlog.

🔗Question 6

🔗Who wrote the copy of the website?

Arthur Mingard: It’s entirely generated with AI

🔗Question 7

🔗Have you hired or considered hiring data scientist to learn from the data captured in telemetry services, QA etc.?

Joseph Denne: We previously had a data scientist who was an actual astrophysicist on the team (they specialised in brown dwarfs if you can believe it). They leant us a lot of insight, and it will probably be useful to get more insight like this in future. But for now, we’re focused on core services and performance.

🔗Question 8

🔗Does the DNS service allow for cnames or aliases at the apex in a similar way to the alias in AWS and if so, does it allow off network?

Arthur Mingard: Yes, this is now supported, both on and off the network. This feature was deployed early last month and has proven to be very useful.

🔗Question 9

🔗How can Edge improve the performance for users connecting using public DNS settings such as 8.8.8.8or 1.1.1.1?

Arthur Mingard: We’re working on a process that removes the reliance on BGP network routing, favoring a more accurate process of locating the closest device. This is likely to be delivered in Q1 2020.

Joseph Denne: BGP requires the maintenance and continual monitoring of routing peers to work efficiently in a network like Edge, and given that the network has a clear and up to date understanding of its own shape, it makes sense for us to look to how we can improve traffic distribution at this level.

🔗Question 10

🔗The list of hosts is showing which Stargate they are connected to. Shouldn’t there be a reference to a Gateway instead? I thought the communication was like this: Stargate < > Gateway < > Host. Do Hosts skip going through a Gateway if a Stargate is closer?

Arthur Mingard: Stargates and Gateways serve two very different purposes. Stargates handle DNS, network coordination and service distribution, and Gateways handle http requests, queueing and app distribution. We are looking at ways to make Gateway act as a mirror for the Stargate coordination data so there’s every chance this process will eventually add some performance increase. It’s worth noting that once DNS has resolved to a Gateway, the requests and responses are handled by Gateway and Host only, and have no dependency on Stargate other than low loading operations like coordination.

🔗Question 11

🔗When will we be able to downscale the hosts? I would love to have a node on a server that doesn’t meet the requirements of a raspberry but very close.

Arthur Mingard: The Raspberry Pi is a very low powered device so it’s unlikely we’ll be deploying CDN to something with less resources, so your device might not be able to contribute just yet. Storage will work on lower powered devices when we introduce it to the network. We will always aim to make the best use of resources and there is every possibility future service will have even lower requirements.

🔗Question 12

🔗What is the current process for someone to remove their stake and shut off their node? Will there be an easier way in the near future?

Joseph Denne: To shut it off you can just pull the power. To return the stake, right now you need to reach out to an admin, who will then disconnect your machine and give you instructions to prove wallet ownership. Your stake will then be returned. (Note that there is a 30-day cooling off period for Hosts.) And yes, we’re working on automating this whole process.

🔗Question 13

🔗Can marketing materials be made available to the community to assist with referrals to the network? Items such as downloadable brochures, flyers, post cards, etc. Also, who do you envision as the target market in these early stages?

Arthur Mingard: Throwaway culture is something we should all be keen to avoid. A business card or leaflet might seem insignificant, but with such a high percentage of them being discarded within hours of being received its more important than ever to think carefully before contributing in this way. Digital resources are the way forward, and that’s something we are keen to improve on!

Our target market - anything from a personal site to an enterprise platform. The concept of lowering cost and simultaneously increasing availability at the edge is applicable across the entire spectrum of digital propositions.

🔗Question 14

🔗Will Edge be listing on node monitoring sites such as Masternodes Online? If so, which ones and when?

Joseph Denne: Good shout. Yes. We’ve spoken to a few of these sites in the past and will pick up the conversations again now self-onboarding is live.

🔗Question 15

🔗The number of Stargates has gone down to 7, is it downsized?

Arthur Mingard: To improve network performance and reduce the number of hops we have removed Stargates that were not receiving traffic. These idle Stargates were serving no traffic, bouncing DNS requests on to other Stargates which was having an adverse effect on the speed of DNS resolution. We anticipate introducing more Stargates soon, in areas where there is higher demand.

🔗Question 16

🔗When will there be GPU-enabled nodes?

Joseph Denne: We’re working on giving the CDN application access to the GPU units in devices, which will be the first practical application of GPU resources within the network. Specifically, it will be used to increase the performance of image manipulation functions as part of the media pipeline in the Content Delivery service.

🔗Question 17

🔗When will Edge have onboarding from the site? It’s a bit weird to have to contact admins on Telegram for it.

Arthur Mingard: The initial phase of self-onboarding was designed to give us the flexibility needed to plan UX of the next iteration without introducing any technical debt. We are now moving onto the second phase, where an automated version will replace the current manual validation.

🔗Question 18

🔗I’m curious about the host SD card capacity cost/benefit as well. I planned on using a 128gb card on my Pi 4 but it was actually corrupt on delivery so I used an available 64gb. I figure that there would be, presently, no benefit but perhaps some future proofing for eventual features is worth planning for?

Arthur Mingard: Firstly, sorry to hear about your SD card. That’s very frustrating! Absolutely a good idea to offer up more space on the device - it will in time be used and compensated for.

🔗Question 19

🔗Why doesn’t my FN have the same behavior as my self-onboarded host? Because my FN works perfectly fine. They are connected to the same router & power outlet. They should have the same session lengths. Do they use different software or calculations on your back-end?

Joseph Denne: Each node on the network will perform different jobs depending on demand (jobs in the Gateway’s queue), latency, existing work done, and a few other variables. No two nodes are the same, no two nodes perform the same work, therefore no two nodes will earn the same amount.

🔗Question 20

🔗How about putting a QR code on the Edge dashboard for the staking address?

Arthur Mingard: This could be useful for improving mobile usability. We’ll put this in the backlog.

🔗Question 21

🔗My self-onboarded node has been online for all days except for a week early on in Sept. However, the traffic seems very unstable. How can it be that some days there is zero revenue, not 0.1 or 0.5 or something like that but 0? What kind of sites are currently running Edge services in the network? To me it feels like the algorithm for calculating earnings must be broken as I can’t see how the earnings pattern can reflect traffic on normal sites.

Arthur Mingard: Some sites on the network have a long cache, often between 1 day and 1 month. This level of caching means that there can be days where the number of requests are lower. If there are a number of Hosts connected to the same Gateway as your Host and perhaps they have a better connection, they will resolve more requests than yours. If there are very few requests then it is possible that your Host might not be used heavily. You are still payed a base value for the Host.

🔗Question 22

🔗Can we get clarification on what Hosts are currently doing in the network? Are they sending images/streaming video directly to the end user browsing a webpage that’s hosted on the network or do they have to route their traffic through a Stargate before getting to the end user?

Joseph Denne: In short, Hosts route traffic through Gateways: there’s no one to one relationship between an end user and a host, and their existing is completely hidden to the end user. Hosts are currently processing media (resizing, scaling, cropping etc.) ahead of delivery to the end user as part of the Content Delivery service. They also cache locally. Soon they will be used for storage, and in time for processing end customer apps as well.

🔗Question 23

🔗How will a trusted execution environment work with online software that’s based on JavaScript?

Arthur Mingard: The network isn’t written in JS - it’s in Go. And it provides an architecture that abstracts storage and processing from the online environment.

🔗Question 24:

🔗Why are a number of Hosts showing offline but docker ps shows containers running?

Arthur Mingard: We are addressing an issue with cached sessions now showing. No need to restart your device, but if you do see that docker has been running for more than 30 minutes and your device doesn’t show as online do get in contact and we’ll address it.

🔗Question 25

🔗Edge wants to be a central player in the edge computing market, but seems somewhat isolated in the industry right now. Are there plans on joining industry consortiums like LFEdge (Linux Foundation) or Edge Computing (Industrial Internet Consortium)? Is Edge striving for collabs with larger players in the cloud/edge market?

Joseph Denne: This is something that we have an eye on, but we have to pick what we focus on as there’s a lot to do, and limited resources. And while this sort of thing can be useful for pushing the industry forward and exposure, they can also use up a lot of time to little material effect. So in short, we have an eye on the various consortiums that are out there, and we’ll approach them when the time is right.

🔗Question 26

🔗What’s the current relationship between EDGE and INDX capital? Is Edge keeping an eye on company practices for those listed as ‘partners’? Was there a determined timeline in the partnership agreement for them to onboard their nodes in the network? How are they fulfilling their end of the deal?

Joseph Denne: We know the guys - they are here in London - and have leant them support. They tested with our containerized service and as far as we’re aware, the intention is for it to be used in time. We keep up to date on an ad-hoc basis, but haven’t spoken to INDX for around 3-4 months as far I’m aware.

🔗Question 27

🔗Can you elaborate on the current way of locking stakes in the originating contact and how this will be done in the future? When can we expect trustless (smart-contract-based) locking of stakes on self-onboarding?

Joseph Denne: It’s currently managed by the team as an interim measure, with funds going to a dedicated staking wallet, and from there to the originating contract. This is for ease of audit. In time we’ll be moving to a contract-based, fully automated process.

🔗Question 28

🔗Is hosting large scale ML or astronomy datasets a viable use case for the upcoming file-storage?

Arthur Mingard: Object Storage is meant for static files rather than evolving data, but there’s no reason that it couldn’t be used for this sort of data. We’re moving to Beta in the coming months for ephemeral data and to persistent data thereafter. The beta period will give us a lot of data/feedback which will in turn impact the use cases for the tech.

🔗Question 29

🔗Could peer2peer syncing protocols (e.g. bittorrent) be used to downloads large amounts of data directly from the hosts?

Arthur Mingard: There’s no plan for this at the moment, but the idea is sound. We’ll give this some thought.

🔗Question 30

🔗Have you considered deploying an easy to understand privacy policy framework for Edge hosts (and network as a whole)? Mainly to gather feedback and provide education on that front too from early on.

Joseph Denne: There’s a bunch of legal work to get through, and this is part of that. The setup of the network is within the legal of the UK, and the security approach is designed with the protection of individual hosts in mind. Anyway, yes!

Related articles

More knowledge articles
Mail icon

Like what you are reading? Want to earn tokens by becoming an Edge Node? Save money on your hosting services? Build amazing digital products on top of Edge services? Join our mailing list.

To hear about our news, events, products and services, subscribe now. You can also indicate which services you are interested in, which we use for research and to inform the content that we send.

* You can unsubscribe at any time by emailing us at data@edge.network or by clicking on the unsubscribe link which can be found in our emails to you. Read our Privacy Policy.
Winner Best Edge Computing Platform Technology & Innovation Awards 2019
Presented by Juniper Research