Some 10 days in the past, we revealed the primary a part of a Q&A that Piotr, our performing CEO (and all-time CTO) answered for one in all our customers. As promised, there was a component II (and doubtless, we could have a 3rd instalment with follow-up questions).

This second half comes proper on time as we put together for the discharge of Graphene, the LibraryOS we have now been constructing in a working group with Intel, ITL, Chia-Che Tsai and Don Porter. Final week, we revealed an interview with Mona Vij, R&D lead at Intel, that defined the company’s relationship with the venture. On this publish, you possibly can study why we select to go down the Intel SGX route, which resulted in Graphene.

One other fascinating matter Piotr covers in his solutions is token (GNT) and velocity. As you might know, we wish to present avenues that confide in giving GNT extra worth (worth within the context of the perform, and never worth) and the query on token velocity is a wonderful primer to know GNT.

Q: What are your ideas on SGX? Would embracing this expertise imply that Golem service suppliers can be restricted to Intel chips solely?

PJ: Firstly, I want to underline that Golem will not be restricted to Intel CPUs – permit me to clarify additional under:

The invention protocol talked about in reply to query no 2 permits including new varieties of sources. This is applicable to all courses of Trusted Execution Environments (TEEs), together with Intel Software program Guard Extensions (Intel SGX).

We selected to take part within the improvement of Graphene largely as a result of Intel SGX is, by far, probably the most mature one when it comes to provided options and recognizability. The maturity of SGX is important to us as a result of as acknowledged in query no 6, we’re constantly on the lookout for further options to extend the standard of the Golem community (be that computation integrity or confidentiality).

Redundant verification together with Intel SGX might permit the community to succeed in fairly excessive ranges of integrity and confidentiality, even within the presence of many different CPUs.

The opposite TEE options usually are not fairly there but, however by design, they need to be simple so as to add to the community. For now, if requestors want to profit from the extra security measures provided by SGX, they need to have the ability to take action. Sooner or later, selecting different applied sciences ought to be simply as simple.

We have already carried out a working integration of Blender in Graphene for Golem, however each, the combination and the Graphene LibOS itself nonetheless want further improvement work earlier than they’re launched into mainnet – which matches alongside nicely with the reply to the fifth query. The Activity API ought to permit builders to organize their integrations, and they need to be capable of present integrations using many several types of useful resource (e.g., a number of GPU or CPU varieties, together with Intel CPUs with enabled SGX).

The essential level right here is that SGX and TEEs, basically, symbolize courses of expertise centered on what’s broadly understood as trustworthiness. Nevertheless, {hardware} with these options will not be privileged over some other classes of sources in Golem.

For instance, the WASM integration takes benefit of an execution setting that comes with its particular set of options. One other execution setting is QEMU, which we’re observing intently.

Q: The place have you ever obtained to with enterprise partnerships?

PJ: We’re presently specializing in the product and never trying actively for enterprise partnerships. Nevertheless, as acknowledged in reply to the query no 4, there are corporations and people occupied with our expertise, however at this stage, we’re on the lookout for alternatives to ship a greater product reasonably than to ascertain partnerships.

We’re an lively participant within the Graphene Consortium (https://grapheneproject.io/), the place Intel is among the official contributors. For now, evidently Golem is the main drive behind the path and the event of the platform.

Additional information has been revealed in our AMA:

Q: Have you ever been working with any dApp builders to combine Golem into their choices?

PJ: We’ve spent a substantial period of time interviewing potential customers and analyzing the ensuing use instances (and we nonetheless do). Our essential purpose, although, is to know the way in which Golem ought to be offered to the builders to achieve success – not on the lookout for dApp or any utility builders to begin integrating with Golem proper now.

As soon as the Activity API is launched, we anticipate exterior integrations with Golem being born. This API will likely additionally permit customers to make the perfect use out of the Graphene options.

Within the meantime, we plan to launch just a few extra integrations (e.g., the WASM integration permitting for a wide range of apps, and the transcoding integration) to make sure that the platform handles completely different courses of duties accurately.

Q: What are your ideas across the velocity argument with regards to the worth of the token?

I will refer implicitly to GNT which is an instance of a utility token, not an arbitrary one. Token Utility is of prime significance.

If the only real function of the token on a platform is to be the medium of change, then favouring this token over some other, with comparable traits (e.g., common transaction time, transaction charges denominated in USD, and so forth) will not be mandatory from the platform usability perspective. After all, there may be the recurring argument that the community impact of a venture, which makes use of its medium of change token, is an important issue. Subsequently, forking the venture to make use of some other token will not be going to be any extra worthwhile than utilizing the unique token. This appears rational supplied the community results talked about above successfully exist.

Alternatively, one can think about forking all tasks, which implement their ERC20 as a medium of change solely. For every venture, its native token would get replaced with ETH. From the consumer perspective, it might lead to higher UX, as solely a single token must be used down the highway.

One further rationale to take action can be utilizing a single token, as presently, any ERC20 transaction requires ETH to pay fuel charges in any case. It ought to be an fascinating train to mannequin such a state of affairs and attempt to estimate its impact on the ETH worth basically, and the estimated worth of ETH nonetheless held by each such venture. This simulated state of affairs implicitly assumes that hypothesis is both dominated out from the image or no less than considerably lowered (which isn’t sensible in fact).

Briefly, if a workforce decides to create a separate token for the venture, and the community impact can’t be assured shortly, then the Utility of the token is of prime significance. Nevertheless, evidently presently there may be nonetheless room for added project-specific tokens used solely as a medium of change.

Getting again to your query, it appears you’re referring to Vitalik’s well-known blogpost in regards to the medium of change tokens (https://vitalik.ca/basic/2017/10/17/moe.html). On the one hand, he highlights just a few necessary features of the described token mannequin. Nevertheless, however, evidently a big a part of his reasoning might not essentially be right, as acknowledged within the doc (https://basicattentiontoken.org/token-econ.pdf) written by Scott Locklin.

Lengthy story brief, velocity might affect the token worth each negatively and positively. If the rate is proscribed resulting from technical causes, then it could immediately render the venture unattractive, which can lead to a decline of the token worth. Although, excessive velocity of the token in a state of affairs, the place solely a fraction of tokens is used, might appeal to extra customers to make use of the platform to purchase the token, rising its worth.

These had been some fairly easy and a bit contrived examples, however evidently there isn’t a simple correlation of the rate together with the worth of the token. However, creating incentives to stake the token seems to be an necessary issue on this puzzle.

That is why a token financial layer ought to incentivize customers of the platform to stake, or the venture ought to introduce some sink mechanisms. Relying on the general token mannequin, sensible time-locking could also be sufficient, and burning tokens might not be mandatory.

Furthermore, it is a particular argument for having a separate token as customers would not most likely be so joyful in the event that they needed to stake their ETH to take full benefit of the dApps’ performance particularly as a result of the staking time could also be arbitrarily lengthy.

With the above arguments in thoughts, the ultimate thought is, that crucial factor is to ship an excellent, working product or platform, with nice UX associated to the token in use, and that is what Golem is attempting to do. The intrinsic incentives to stake make sense solely when the product can be utilized. In any other case, it largely boils all the way down to hypothesis.


Thanks for studying, as soon as once more! We extremely encourage everybody to publish the follow-ups to those questions and the earlier publish within the Reddit thread.

We hope you loved this Q&A bit and stay up for your suggestions.

Extra details about Golem on our web site.

Supply hyperlink


bitcoinBitcoin
$
10,586.11
$
10,586.11
0.54%
ethereumEthereum
$
175.03
$
175.03
0.36%
rippleXRP
$
0.256288
$
0.256288
1.14%
bitcoin-cashBitcoin Cash
$
292.20
$
292.20
0.45%
litecoinLitecoin
$
65.49
$
65.49
2.15%
binancecoinBinance Coin
$
23.01
$
23.01
6.86%
tetherTether
$
1.00
$
1.00
0.37%
eosEOS
$
3.26
$
3.26
1.33%
bitcoin-cash-svBitcoin SV
$
130.50
$
130.50
0.53%
cardanoCardano
$
0.044614
$
0.044614
1.4%