stem
for evaluating Ambassadors, but the business logic of such a system
are expected to be more complex and differ more substantially among
participants.
The reference PolySwarm daemon will “decay” old data, maintaining
a “sliding window” into participants’ reputation. We believe a
sliding window approach is necessary to avoid lockout of new market
participants.
We anticipate making use of the following data points in a reference
Ambassador implementation:
p. 23
Version 15
1. Assertion accuracy (relative to ground truth and/or off-chain
analyses).
2. Frequency and average response times for both Bounties and
Offers. In the case of Offers, Ambassadors may choose to publish
aggregate data that is normally only observable within an Offer
Channel. Ambassadors may favor Experts who demonstrate rapid and
consistent response times.
3. Ability to maintain confidentiality of Artifacts delivered under
Offers. This is likely a hybrid on-market, off-market analysis.
Ambassadors may choose to review their competitors’ past Verdicts
and attempt to uncover errors (e.g., Ambassador A stated that
the sample was benign when in fact it was malicious). Experts may
seek to avoid doing business with Ambassadors whose due diligence
fails to hold up to such scrutiny in general or is more frequently
incorrect under specific circumstances (e.g. Ambassador A frequently
miscategorizes Microsoft Word documents resulting in certain Experts
avoiding analyzing Word documents on behalf of Ambassador A).
In other words, among other inputs, Experts may be interested in
Ambassadors’ Verdict accuracy.
Consumers will likely rely on aggregate Consumer Reports or AV
Comparatives-style websites and publications that report on each
Ambassador’s track record, perhaps with sample inputs with known
malintent and verification that an Ambassador’s Verdicts always
match what it represents back to its clients.
Worker Registry
Experts in the PolySwarm ecosystem will, in many cases, choose to
encapsulate their expertise into automated Workers . Workers are systems
that automatically respond to Bounties and Offers, perhaps with expertise for
certain filetypes or network traffic . The PolySwarm marketplace will provide
a distributed Worker Registry to ease discovery of Experts’ offerings .
We will develop a common interface and description language
for publicly registering Workers. These components will interact
with a PolySwarm Registry that is modeled after initial Ethereum
dApp registry smart contracts. PolySwarm’s Worker Registry will
contain descriptions that advertise particular skillsets (e.g., “I’m
good with Mac OS Mach-O binaries” or “I handle Microsoft Word
Documents”).
p. 24
Version 15
Artifact Confidentiality
Confidential disclosure of Artifacts is a first class feature in the PolySwarm
marketplace via the Offer instrument . Nevertheless, without taking additional
precautions, Ambassadors utilizing Offers must trust Security Experts’
motivations and technical ability to ward off attackers, possibly in perpetuity .
We expect this trust requirement to give rise to secondary markets
and novel applications of technology with the goal of minimizing
or consolidating trust. The following are several examples
of instruments built on top of PolySwarm to further address
confidentiality requirements.
Trust Minimization via Intel SGX
Intel SGX is a new technology initially shipped with the Skylake family
of processors. At a high level, SGX provides opaque computation
environments inside “enclaves”. Memory contents of SGX enclaves
are unavailable to other software and users on the same system,
irrespective of permission level. In the context of PolySwarm, SGX
enclaves would allow Experts to scan Artifacts without revealing
the Artifacts to the Researcher and without revealing the Experts’
intelligence to anyone else. SGX therefore promises mutual trust
reduction between Ambassadors and Experts.
Experts may choose to run accredited (and attested) SGX enclaves
that ingest the Artifact and their intelligence and produce an
Assertion that may be returned to the Ambassador. There are
technical hurdles to overcome, but we are confident SGX can be
integrated and scaled with enclaves bootstrapping one another in
a distributed fashion. Future iterations of PolySwarm may allow
Ambassadors to author future “SGX Bounty” statements into the
PolySwarm market, revealing the Artifacts only to bootstrapped
enclaves.
Trust Consolidation via Scan-as-a-Service (SaaS) Ambassadors
Experts may choose to outsource computation tasks to third parties
for a number of reasons that may or may not be PolySwarm-specific.
For example, we expect many Experts to make use of AWS, Google
Compute Engine, Azure or other such on-demand computational
services to scale their Assertion pipeline in response to PolySwarm
market demand. This is not an intrinsically PolySwarm-motivated
compute decision.
p. 25
Version 15
We expect PolySwarm-specific incentives for SaaS entities to arise
naturally. In the context of confidentiality, participants will create
SaaS companies that serve as trusted third parties sitting between
Ambassadors and Experts and maintaining nondisclosure agreements
with both. These SaaS companies would serve the same purpose as
SGX enclaves - protecting Artifacts from Experts and intelligence
details from Ambassadors.
Ambassadors may find it easier to trust third party SaaS entities with
public reputation rather than a large number of individual Experts whose
identity may not be known. Such Ambassadors will author more Offers
targeted at SaaS entities rather than directly at Experts.
Additional Markets
The PolySwarm market will create demand for a number of
additional markets . One such market (Scan-as-a-Service)
was previously discussed in the context enabling additional
confidentiality controls . Below are some additional expected
markets .
Ambassadors
This is an obvious market and one heavily discussed throughout this
document. Many Consumers will choose to do business with the market
through an Ambassador, but working through an Ambassador is not a
strict requirement.
Artifact Publishing Services
When placing a Bounty or Offer, the Ambassador must make their
Artifact available to Experts. Artifact hosting entities will offer
this service, making such sharing seamless. Their clients will be
Consumers and Ambassadors. We expect many Ambassadors to also
offer Artifact Publishing services.
Reputation Tracking Services
In the United States, three credit bureaus are trusted by banks to
maintain a ledger of “facts” regarding individual trustworthiness. The
PolySwarm market will present a similar demand; market participants
will want to quantify the trustworthiness of other participants for
a variety of reasons already discussed. It is likely that reputation
services will arise that track trustworthiness of participants and offer
this information to mitigate counterparty risk.
p. 26
Version 15
Roadmap
Here we present preliminary PolySwarm development milestones . Each
milestone will: (1) offer new functionality and (2) provide documentation and
tests that enable new Ambassador, Expert, and End User transactions .
PolySwarm is at a distinct developmental advantage relative to
networks like Bitcoin and Ethereum that must retain all transactions
in perpetuity. Once ground truth is determined or an Offer Channel
is settled to the chain, Ambassadors and Experts have received what
they need out of that transaction: intelligence regarding the Artifact
and NCT, respectively. This atomicity provides leeway and allows us
to make backwards-incompatible changes, up until the 1.0 release,
without cost and consensus challenges experienced in other networks.
V0 .1 Alpha (30 April 2018)
PolySwarm Alpha will focus on delivering a prototype end-to-end
workflow for Bounties after successful token sale. This initial
release will allow the PolySwarm team to begin the bootstrapping
process in earnest. Ambassadors and Experts will be able to transact
via a PolySwarm test network (“testnet”). The basic Bounty smart
contracts will be available on the PolySwarm testnet, allowing
Ambassadors to place Bounties, Experts to render Assertions, and
Ambassadors to render Verdicts.
We anticipate that PolySwarm Alpha will accomplish the following:
• A “Bounty Manager” smart contract which holds Bounty Amounts and
Fees, accepts Assertions, publishes Verdicts, and rewards accurate
Experts
• Artifact posting and retrieval to / from off-PolySwarm storage
(e.g. S3 and/or IPFS)
• Prototype algorithm for selecting Arbiters at Bounty expiration
• Reference implementations for Bounty publishing, Assertion response,
and Arbiter ground truth determination
V0 .2 Beta (31 May 2018)
PolySwarm Beta will focus on delivering PolySwarm Offer support over
Raiden-style Channels. Offers are designed to facilitate higher-
throughput Artifact evaluation relative to Bounties and as such
demand more efficient blockchain settlement for each transaction.
PolySwarm Beta will implement Offer smart contracts. We anticipate
implementing this functionality as direct, peer to peer, Raiden-style
Channels established between Ambassadors and Experts which utilize
the testnet.
p. 27
Version 15
We anticipate PolySwarm Beta will accomplish the following:
• Offer Channels
• Distributed Channel establishment
• multiple Offers via a single Channel
• NCT settlement at Channel Close
• Reference implementation for Offer production and response
• Reference implementation for Arbiter ground truth notification and
response
• Fee collection and distribution smart contracts
• Research and quantify trust (in NCT) required for a specific Artifact
volume
• Offer/Bounty development toolkit: a test and implementation
framework for Experts and Ambassadors to automate servicing of
Bounties and Offers
V0 .3 Gamma (31 July 2018)
PolySwarm Gamma will focus on providing matchmaking features that
facilitate collaboration between Ambassadors, Experts, and Arbiters.
Experts will be able to encapsulate their expertise into Workers tuned
to specific types of Artifact. We expect them to be able to advertise
these Workers in a PolySwarm Registry. This Worker Registry should
allow Ambassadors to engage with Workers in a risk-free manner
prior to the Stable release (Gamma will continue to utilize the
PolySwarm testnet).
Additionally, this release will consider enhancing and extending the
Bounty instrument in light of Alpha and Beta feedback and experience.
As one example, we anticipate that Bounties will benefit from the
ability to solicit Assertions that are kept confidential from other
Experts and revealed only after the Assertion deadline.
We anticipate PolySwarm Gamma will accomplish the following:
• Worker Description Language (WDL) describing artifact analysis
capabilities and Expert authorship attribution (for reputation)
• Distributed, searchable, registry of Workers and their capabilities.
• Confidentiality for pre-deadline Bounty Assertions
• Finalize Arbiter selection, response, and end to end Bounty feedback
loop
We anticipate that Gamma will be feature complete.
p. 28
Version 15
Polyswarm 1 .0 (Q4 2018)
PolySwarm Stable will focus on eliminating bugs in PolySwarm Gamma
prior to migration from the PolySwarm testnet (test tokens) to the
real PolySwarm market.
As a stretch goal and as time / funding allows, Stable may introduce
first-class Artifact confidentiality via a limited-audience Bounty.
Polyswarm 2 .0 (Q2 2019)
PolySwarm’s second stable release currently has two main goals:
expanding Offers and Bounties to handle a wider range of Artifact
types (e.g. URLs and Network streams) and easing PolySwarm
utilization for the Enterprise and Home User.
Artifact type expansion in future PolySwarm releases should make
it possible to handle, for example, automated phishing attempt
detection of end user URLs by placing Bounties and Offers directly on
hosted content.
Extending PolySwarm’s protections directly to the endpoint will be
a priority for post-Stable development. We anticipate developing
an Open Source reference implementation of an Endpoint protection
suite which automatically blocks malicious artifacts based on
PolySwarm’s intelligence. We see this open agent framework as a key
to widespread end user adoption.
Future Features
PolySwarm’s future releases will mirror the evolving threat landscape
and real-world usage of the network. We see several opportunities,
contingent on observed PolySwarm usage, for the following features
to benefit the network and the ultimate goal of end user protection:
• Flexible Artifact confidentiality
• Basic: Artifacts selectively disclosed to Experts or SaaS providers
• Advanced: Artifacts never revealed to anyone outside of SGX
enclaves of fully homomorphic engines
• New Instruments (e.g. subscription-ready threat feeds)
• Deployment support for endpoint-resident Worker containers that
supplant today’s endpoint protection suites.
p. 29
Version 15
Token Sale
Technical development, community engagement and participant outreach
for the PolySwarm ecosystem will be conducted by PolySwarm Pte . Ltd . with
funding derived from the sale of PolySwarm's Nectar utility tokens (NCT) .
These tokens will be exchanged by participants for threat
intelligence. NCT powers all the major PolySwarm interactions:
Bounty placement, Offer channel establishment / teardown /
exchange, assertion registration and ground truth determination. By
selling utility-specific NCT tokens, exchange rates for NCT to threat
intelligence should enjoy isolation from the volatility in the larger
ETH market, ensuring active participation in PolySwarm without
concern for fluctuations in ETH.
This section is intended to give a high level overview of PolySwarm’s
tokenomics. Should any conflict between this document and the Token
Sale Agreement (“Agreement”) arise, the Agreement shall take
precedence. The Token Sale Agreement is available online.
Schedule
PolySwarm Pte. Ltd., Inc is currently operating a token PreSale for
invited participants. The public NCT sale is scheduled to open to
qualified buyers on February 20, 2018 at 19:00 UTC. It is scheduled
to close March 22, 2018 at 19:00 UTC.
Purchase Requirements
NCT will be made available exclusively for purchase with Ether (ETH).
Potential buyers are advised to secure sufficient ETH in advance of
the sale.
The token sale will employ Know Your Customer (KYC) controls with
a unique technical mechanism to tie specific purchases to KYC
data points including, but not limited to: Country of Origin, email
address, full name, IP address and acceptance of the terms in the
Token Sale Agreement.
p. 30
Version 15
Caps / Limits
No new tokens will be created after the close of the token sale
period.
The total number of NCT tokens will be determined during the token
sale.
The maximum number of tokens is capped. This cap is based on the
maximum funding USD cap, the price of NCT (in ETH) and the current
ETH to USD exchange rate. Refer to the token sale web page for
specifics on NCT caps, price, funding tranches and more.
Allocation
70% of NCT tokens will be sold during the Token Sale.
15% of the remaining 30% will be distributed to enterprises, vendors
and Security Experts interested in helping bootstrap the PolySwarm
ecosystem. Recipients will be chosen at PolySwarm Pte. Ltd.. sole
discretion.
The final 15% of tokens will be used by PolySwarm Pte. Ltd. in a
manner that best accelerates adoption of the PolySwarm ecosystem.
As an illustrative example, tokens may be offered as incentive to
participate in PolySwarm “hackathons”. This allocation will be at
PolySwarm Pte. Ltd.. sole discretion.
p. 31
Version 15
Distribution
We envision that ETH derived from the sale of PolySwarm's Nectar
utility tokens (“NCT”) will be allocated in the following manner:
54.51% - Protocol and Software Development
3.00% - Office Expenses
4.06% - Operational Overhead (travel, hosting, etc)
14.49% - Legal
13.62% - Marketing
10.32% - Tax
The PolySwarm team expects Legal spending to decrease in FY19/20.
The surplus will be redistributed into Marketing, Overhead and
Developer Salaries.
Our 10.32% estimated Tax is low, but not unreasonable. This is
good for participants: more funding will be allocated directly to the
development and promotion of PolySwarm.
14.49%
Legal
Marketing
13.62%
Tax
10.32%
3%
Office Expenses
Operational Overhead
Protocol &
Software Development
54.51%
p. 32
Version 15
Disclaimer
This whitepaper is intended to provide technical background
to would-be token purchasers . The PolySwarm token sale will
fund initial prototype development and testing of the economic
instruments presented here . We expect, as would any prudent
participant, that the details presented in this document may
change during pre-Stable development and testing periods .
PolySwarm Pte . Ltd . and token purchasers’ interests are
aligned to make PolySwarm a viable marketplace that truly
disrupts the threat intelligence industry .
Conclusion
PolySwarm is a novel market and associated ERC20 token,
NCT , that is posed to disrupt the threat intelligence industry
by facilitating new methods of interaction between security
experts, vendors and enterprises . The PolySwarm team will be
conducting a token launch to fund software development and
to cultivate a community around this new marketplace .
We hope you’ll join us in protecting users from All Angles™ .
p. 33
Version 15
polyswarm.io 2018 © PolySwarm Pte. Ltd.프로젝트로 돌아가기