← All themes

Cloud Decentralization

22 statements · 2020–2020

60% of Ethereum nodes run on AWS — that's not decentralization

"Hashing It Out #70 - Akash Network - Greg Osuri" (Hashing It Out (The Bitcoin Podcast Network))

“60% of Ethereum nodes are running on AWS right now, which is kind of sad… it doesn’t really matter how decentralized the network is, [it] really comes down to your physical deployment. If you’re running on Amazon, I don’t consider that decentralized at all, especially when the majority of your network is running on a single cloud provider.” — 00:58:06

Context: His motivation for DISCO (decentralized infrastructure for serverless computing operations) and home nodes (quote spans into the 00:58:49 block); he adds that with physical access and no SGX, an adversary intending to shut Ethereum down could do it “in a matter of minutes” (00:59:31).


Akash gives placement sovereignty to the user — and clouds themselves plug in

"Hashing It Out #70 - Akash Network - Greg Osuri" (Hashing It Out (The Bitcoin Podcast Network))

“What Akash does really is open up the market to these massive clusters, to small little deployments all over the world and gives the sovereignty to the user to choose what they want. And of course, the multi-cloud architecture in Akash also means that cloud companies plug into Akash directly, which they’re doing right now.” — 00:19:50

Context: From the recovered mid-episode span: latency-sensitive jobs can pin local-residency clusters, then “switch over to a more cost optimal deployment” (00:20:36); he says tier-2 cloud providers were already plugging in, framing Akash as additive multi-cloud rather than a cloud replacement.


150 edge metros at sub-10ms latency via the Packet program

The Akashian Challenge Phase 1 Livestream (Akash Network)

“We have about 150 edge sites that Akash is ready to go on… when we unlock those edge sites through Packet, I believe it’s going to be 150 metro areas delivering at under 10 millisecond latency — that’s extremely attractive for high-content, high-bandwidth streaming.” — 00:50:19

Context: On the Packet (Equinix) edge program; positioning Akash for streaming and low-latency edge workloads circa 2020.


Interoperability is the future of the decentralized cloud

The Akashian Challenge Phase 1 Livestream (Akash Network)

“We believe the future is interoperability, because Akash cannot do everything you need for a full cloud deployment… we’d rather work with teams or projects that have their expertise… interoperability, I believe, will be the future and is key for scalability and adoption.” — 00:40:56

Context: On IBC integrations planned with NuCypher (key management), Solana (smart contracts), stablecoins for settlement, and storage networks (IPFS, Sia).


Only 9 of 91 testnet nodes were on AWS, vs Ethereum's 60%+

The Akashian Challenge Phase 1 Livestream (Akash Network)

“There were, I believe, nine or ten nodes that went down right around the time there was an outage in AWS — so only nine out of 91 nodes are actually on AWS. I take that with a lot of pride, because Ethereum has over 60% of the nodes on AWS, or cloud in general.” — 00:06:43

Context: Recapping the AWS outage that hit mid-testnet; used as evidence Akash’s validator base practices the decentralization it preaches.


Unstoppable, borderless workloads as a core use case

The Akashian Challenge Phase 1 Livestream (Akash Network)

“Because Akash is unstoppable — that means you cannot take down the workload unless you have the key to the workload… borderless knowledge sharing, borderless commerce, truly borderless that cannot be stopped, is a big use case for Akash.” — 00:52:25

Context: Describing Phase 3 incentives for “ownerless oracles” (e.g., DAO-safe frontends like Uniswap’s) and censorship-resistant publishing.


85% of global cloud capacity sits unused

"The Akashian Challenge Livestream: Phase 3 Week 1" (Akash Network)

“We tap into world’s unused cloud capacity — 85% of the global cloud capacity that’s sitting in data centers is not used… probably even more, we’re being conservative with our estimates.” — 00:32:55

Context: His supply-side thesis, illustrated with Intuit: TurboTax servers peak at 95-96% in tax season but run ~10% the other nine months, and servers have a ~3-year shelf life — “if you don’t use it you lose it.”


Akash: 2x-10x cheaper than the incumbent cloud

"The Akashian Challenge Livestream: Phase 3 Week 1" (Akash Network)

“Akash Network is the world’s first non-custodial cloud that can offer anywhere up to two times to ten times lower cost than the current cloud.” — 00:30:01

Context: Rebutting a questioner calling Akash a “blockchain solution to a problem that doesn’t exist”; he adds cost savings matter most for machine learning and big-data workloads.


Akash as the gateway to the decentralized cloud stack

"The Akashian Challenge Livestream: Phase 3 Week 1" (Akash Network)

“A big vision for us for Akash is to become the gateway for decentralized cloud… what we want is to enable Akash users to use key management on NuCypher and pay using Akash token… so that cross-chain interoperability is very, very important for Akash.” — 01:00:30

Context: Answering why Akash integrated Kava/USDX; multi-currency settlement so one token pays for compute, storage, DNS, and key management across networks — “just like you don’t use 100 credit cards.”


A world where the cloud runs on Superminis in people's homes

Greg Osuri - AKASH Network (At Stake)

“Building a free and open cloud is very, very important. And extending that to regions that experience oppression is very, very important… So you can imagine a world where people host their own super minis in different parts of the world and run the cloud on those super minis. That’s a fantastic outcome, right?” — 00:22:57

Context: Answering a question about censorship resistance; ties home compute to unstoppable, censorship-resistant services (messaging in oppressed countries) and to peers like Orchid and Sentinel.


Akash as the gateway to a "Cambrian explosion" of decentralized cloud

Greg Osuri - AKASH Network (At Stake)

“We envision Akash to be in that forefront as a gateway to the decentralized cloud… I think we’re just seeing a Cambrian explosion of decentralized dCloud or dWeb… of projects that are actually shipping products today… the next step for us as an industry is to figure out how to reduce friction for adoption.” — 00:33:23

Context: Vision of Akash orchestrating Handshake (DNS), NuCypher (key management), Filecoin (storage) behind one interface with unified payment — “hyper-connected and hyper-interoperable decentralized cloud.”


Akash's mission is to break the cloud oligopoly

Greg Osuri - AKASH Network (At Stake)

“Cloud infrastructure, the backbone of the internet, it’s getting increasingly centralized. It’s controlled by the monopoly of literally four companies… you pay the cloud tax, the Amazon tax, we call it, invisibly. And the mission of Akash is to really break the logopoly.” — 00:05:16

Context: Opening framing of Akash; “logopoly” is Whisper’s rendering of “oligopoly.” Continues into the peer-to-peer marketplace pitch for “resilient and unstoppable applications at a much lower cost.”


Non-custodial UX: a "Uniswap for cloud deployments"

Greg Osuri - AKASH Network (At Stake)

“That is a power of non-custodial interfaces… extending that similar experience for cloud is amazing… We don’t have a web interface yet, but you can kind of see how that will turn out to be like a Uniswap for like cloud deployments.” — 00:13:13

Context: He calls user experience the “unrealized benefit” of decentralized cloud, bigger than the cost advantage — a recurring UX-first argument.


Akash as the gateway to a decentralized cloud

"The Akashian Challenge: Closing Live Stream" (Akash Network)

“We are a big contributor to IBC, and the reason why we wanted to use IBC is because we believe in a frictionless interoperable cloud, [with] Akash really being positioned as a gateway to a decentralized cloud.” — 00:50:16

Context: Greg outlining 2021 plans (mainnet 2, economics/incentives, interoperability) — early statement of the recurring “gateway/supercloud” framing.


Akash as the open-source AWS with provider add-on services

"The Akashian Challenge: Closing Live Stream" (Akash Network)

“Amazon AWS started with S3 and maybe some compute, and today it’s an entire jungle of services — it’s a monolith, they control the whole thing. We see this as the open-source version of that, and we see it going to having that amount of diversity of services available on the network.” — 00:57:20

Context: Answer on private Docker image support, envisioning provider-offered add-on services (log retention, fast storage). Speaker: Adam Bozanich (CTO), not Greg; Greg adds the design principle of “doing less things but doing it really well.”


Akash could host 200 Solanas; supply is not the problem

The Akashian Challenge Livestream: Phase 3 Week 2 (Akash Network)

“100%. I mean, I think we can cover at least 200 Solanas with the current capacity that we have… we don’t have a supply problem.” — 00:49:27

Context: Asked whether rented Akash capacity would suffice in 2021-22 to cover a significant part of Solana’s compute needs; Greg says demand, not supply, is the constraint.


Akash in 2031: a $300B market, everywhere like the cloud

"The Akashian Challenge: Closing Live Stream" (Akash Network)

“300 billion dollars at 100 [percent] market penetration would be nice. It’ll be like the clouds — it’ll be everywhere.” — 00:58:45

Context: Viewer asked what Akash will look like in 2031. Speaker: likely Greg (captions not diarized; Adam follows with “we’ve been on this thing for quite a long time… it could really be a game changer”).


At maturity Akash will have thousands, if not millions, of providers

The Akashian Challenge Livestream: Phase 3 Week 2 (Akash Network)

“How do you detect this illegal content? How do you remove it? Especially considering if you’re Amazon you probably have one company, but in Akash you have thousands if not millions of providers when we reach maturity.” — 00:52:18

Context: Answering how Akash handles illegal content; the scale forecast of “millions of providers” is the notable quantified piece.


Governance-based content takedowns would mean censorship

The Akashian Challenge Livestream: Phase 3 Week 2 (Akash Network)

“The moment you say we can introduce a governance process, that means… you’re introducing censorship… we don’t have the tools and technologies to even understand what the content is.” — 00:53:00

Context: Greg’s contrarian stance that Akash will not adopt on-chain governance for content removal, shifting moderation liability to individual providers instead.


The real problem: cheaper, better, faster compute

"The Akashian Challenge: Closing Live Stream" (Akash Network)

“The real problems we’re solving is cheaper, better, faster compute… Akash, by outsourcing and really bringing on different providers and letting users set the price, you are actually reducing the cost of compute… the amount of steps you need to take to go deploy something on Amazon is exponentially greater than what you can do on Akash.” — 01:12:58

Context: Greg’s answer to “many blockchains aren’t solving real problems — what problem is Akash solving?”; adds “the product is the best due diligence.”


Usability plus censorship resistance is the winning combo

The Akashian Challenge Livestream: Phase 3 Week 2 (Akash Network)

“From one hand you have this incredible usability gains, from the other hand you have this amazing censorship resistance in a permissionless way… all this great adoption that we had is just a testament to these two incredible important features that Akash can provide.” — 00:45:54

Context: After analogizing Uniswap’s few-click swaps vs. a centralized exchange’s 20-step flow, Greg claims Akash extends that same usability leap to cloud deployment vs. AWS.


Users will self-deploy DEX UIs signed by web wallets

The Akashian Challenge Livestream: Phase 3 Week 2 (Akash Network)

“Right now it’s command line, but soon enough… you can see a world where their web interfaces are much more simple interfaces, ideally signed by web wallets, and the users themselves can deploy any DEX UI. So that’s really where we’re going.” — 00:43:48

Context: Discussing ThorChain wanting users to host their own DEX UIs on Akash; Greg frames non-custodial hosting as shifting regulatory liability from creator to user.