Professional Documents
Culture Documents
Key Words:
Abstract
Met averse Pillars of Creat ion (MPC) is t he next major mainnet upgrade t o Met averse
since t he release of SuperNova in 2018. MPC support s a t hree-pronged hybrid
consensus mechanism composed of PoW+PoS+DPoS. T his hybrid consensus algorit hm
will prevent any 51% at t acks, PoS Not hing-at -St ake at t acks, and allow f or Met averse
Smart T oken (MST ) mining. MPC also int roduces t he Met averse Avat ar Reput at ion
Syst em (MARS), an open, decent ralized social credit syst em based on Met averse Digit al
Ident it y. In t he init ial whit epaper draf t , we described a t ype of programmable smart
asset s. In MPC t hese will be implement ed based on verifiable smart cont ract t emplat es.
While designing t he above f unct ions, we lay t he f oundat ion f or a layer-2 archit ect ure in
Met averse called Binary-Port -Chain. T he second layer chain f or st andardized digit al
ident it y and f unct ional smart cont ract s will help ent erprises connect highly scalable
services t o t he main Met averse chain. We call t his dual chain st ruct ure t he Met averse
Binary Syst em.
Foreword
1. First Release (February 2017 t o June 2018): issued ET P and provided t he basic
f unct ionalit y of digit al asset s
2. SuperNova (June 2018 t o March 2019): upgraded t he capabilit ies of digit al asset s
and added digit al ident it y
3. Pillars of Creat ion (March 2019 ~ 2020): dramat ically increase T PS, upgrade digit al
ident it y, and add smart asset s
4. Galaxy (~2020 and on): micro-inflat ion macroeconomic model, orient ed t o blockchain
dat a, will provide digit al ident it y and smart asset st andardized service prot ocols f or
art ificial int elligence.
Overview
Pure Proof -of -Work (PoW) or Proof -of -St ake (PoS) consensus algorit hms have
limit ed t ransact ion speed (T PS) and do not meet t he scalabilit y requirement s of
mainst ream applicat ions. Blockchain designs addressing scalabilit y issues, on t he ot her
hand, are f orced t o make t radeoff s in t erms of securit y vulnerabilit ies. As f ar as t he
current blockchain indust ry scaling plans are concerned, t here are t wo main approaches:
1. Modif y t he consensus algorit hm it self t o improve T PS; or,
2. Modif y t he st ruct ure of t he t ransact ions t o improve T PS.
Met hod 1 is more f oundat ional t han met hod 2 and should be t he pref erred solut ion.
T hus, in MPC we consider a hybrid consensus mechanism t o improve T PS wit hout
compromising securit y. Since Met averse already operat es on PoW consensus, migrat ion
t o hybrid PoW-PoS consensus in MPC is relat ively st raight f orward and f ully compat ible
wit h t he current PoW mining regime.
Common hybrid consensus mechanisms can be divided int o t wo cat egories: PoW+PoS
and PoW+BFT .
Alt hough t he PoW algorit hm is relat ively simple and eff ect ive, it s securit y is a f unct ion
of hash power, raising t he possibilit y of 51% at t acks, selfish mining, and ot her relat ed
issues. By analyzing PoS, we know t hat we can choose PoW+PoS t o compet e f or t he
block generat ion based on t he f oundat ional and f ully compat ible PoW consensus
algorit hm. Finally, t he MARS syst em can act ivat e DPoS on t he Met averse chain. We
divide t he hybrid consensus Pillars int o t wo phases.
Pillars Phase 1
Act ivat ion Point - block height 1924000
Average Block Generat ion T ime is 25.02 seconds, PoW percent age y = 90%, PoS
percent age z = 10%
Af t er calculat ion, due t o t he block generat ion, PoW miners will increase t heir revenue by
about 7% compared t o t he upgrade.
Designing PoS
Pillars Phase 2
Probable act ivat ion block - 2541142
Average block creat ion t ime 16.35 seconds, PoW rat io y = 15/32, PoS rat io z =
1/16, DPoS rat io v = 15/32
Object ive:
1. T hrough t he model design, allow t he DPoS consensus t o be more st able in
generat ing blocks.
2. T he MARS score is t he core evaluat ion of t he block, considering t he qualit y of t he
nodes. Whet her t he node has a cert ificat e, whet her t he st aked amount is sufficient ,
and whet her t he node is act ively generat ing blocks are t hree aspect s f or evaluat ion.
T he cert ificat e reflect s t he management efficiency, normalcy, and guarant eed qualit y
of t he f oundat ion. T he st ake reflect s t he influence of nodes on t he net work, and
how act ively blocks generat e reflect s whet her t he node has f ulfilled it s due
obligat ions during t he operat ion of t he net work.
3. At t he beginning of each node replacement cycle, t here is also a node management
incent ive f or t he development of t he best nodes wit hin t he net work t hat will t ake
int o considerat ion t he number of nodes.
Sufficient condit ions f or act ivat ion: Develop at least 23 secondary nodes
Necessary condit ions f or inact ivat ion: Inact ivat ed relat ive t o 2 million blocks
Not t ransf erable in inact ive st at e
Not t ransf erable in act ive st at e
Cannot t ake t he init iat ive t o cancel t he deact ivat ion
T able 1
Tier 1 Node 23 1.) Hold a Level 1 certificate (valid 1.) After obtaining the first level
for two months) certificate, develop at least 23
2.) Generate up to 46 secondary secondary nodes to participate in the
certificates consensus and activate the validity of
3.) Increase the MARS value after the certificate.
activating the level 1 certificate, 2.) Actively participate in the block
valid for 2 million blocks creation, maintain the MARS value,
4.) Get block rewards based on the avoid freezing that will lead to the
block creation situation revocation of the certificate
Tier 2 Node At most 1.) Hold a Level 2 certificate and 1.) Actively participate in the block
1058 increase the MARS value, valid for creation and maintain the MARS value
2 million blocks
2.) Get block rewards based on the
block creation situation
(1) Hold a valid level one or level t wo cert ificat e, and t he cert score is 30 point s.
(2) St ake number score: For example, 100,000 ET P corresponds t o 30.9 point s, and t he
f ormula is as f ollows:
Among t hem, NodeBlockMiss is t he block error rat e of t he Node f rom t he first t ime it
has become a candidat e. T he recursive calculat ion f ormula is as f ollows, and defines
NodeBlockMiss(0, BlockCount )=0:
3.) MARS relat ed incent ives, as well as node development incent ives (draf t )
The following are projections and are not final figures. These are subject to change as we
garner feedback from the community before MPC Phase II.
T able 2
Basic Flow:
Wit ness-MARS combines t he Sat oshi Nakamot o's random number algorit hm, and t he
MARS score as a weight aff ect s t he probabilit y of select ion. If 1024 is insufficient , it
will be select ed according t o t he act ual person. If more t han 1024, t hen according t o
t he MARS ranking select ed pre 1024, t he MARS influence f act or is degraded t o PoS
only when it is st aked.
Competitive PoW-PoS hybrid block production
First , due t o similar block st ruct ure and consist ent block generat ion logic, we consider
t he simult aneous act ivat ion of PoW and PoS consensus. T he diff erence bet ween t he
t wo is t hat PoS has more coinst ake st ruct ure t han PoW, t hus we only need t o
simult aneously validat e PoS blocks and PoW blocks in t he verificat ion part of t he
consensus.
At t he same t ime, in order t o prevent ident it y f orging and block t ampering, t he PoS
block header also cont ains t he signat ure blocksig of t he ent ire block wit h t he wit ness's
privat e key.
In order t o collect small UT XO and split large UT XO f or mining, each coinst ake
t ransact ion includes a small_collect and large_split f unct ion.
PoS also has a MARS score requirement t o f ulfill bef ore mining.
At t he same t ime, in order t o prevent ident it y f orgery and block t ampering, t he PoS
block header also cont ains t he signat ure blocksig of t he ent ire block wit h t he wit ness's
privat e key.
DPoS act ivat ion requires meet ing t he f ollowing t wo condit ions:
1. Init iat e a t ransact ion t o regist er as a wit ness candidat e;
2. Lock a cert ain amount of ET P as st ake.
Advantages of Metaverse Hybrid Consensus
PoS Analysis
A discussion of MPC design advant ages begins win an analysis of current PoS
variat ions, such as Pure PoS1.0/PoS2.0/PoS3.0, Dynamic PoS, Liquid PoS, Lease PoS,
and Forging PoS. All PoS implement at ions are based on t he f ollowing logic:
On t he PoS prot ocol, blocks are separat ed int o t wo dist inct t ypes: PoW blocks and
PoS blocks. T he PoS in t he new t ype of blocks is a special t ransact ion called
coinst ake (named af t er PoW special t ransact ion coinbase). In t he coinst ake
t ransact ion, t he block owner pays himself t hereby consuming his coin (or coin age),
while gaining t he privilege of generat ing a block f or t he net work and mint ing f or PoS.
T he first input of coinst ake is called kernel and is required t o meet a specific hash
t arget prot ocol, t hus making t he generat ion of PoS blocks a st ochast ic process
similar t o PoW blocks. However, a significant diff erence is t hat t he hashing operat ion
is done over a limit ed search space (more specifically one hash per unspent wallet -
out put per second) inst ead of an unlimit ed search space as in PoW. T hus, no
significant consumpt ion of energy is involved (King & Nadal, 2012).
In ot her words, PoS it self cont ains PoW, and in t he Pure PoS algorit hm, t he PoS part is
st rengt hened and t he PoW part is weakened. We analyzed PPCoin, NovaCoin, YaCoin,
and BlackCoin, and f ound t hat t he PoW+PoS hybrid mode was adopt ed early on.
T o generat e blocks on t he PoS algorit hm, t he f ollowing condit ions must be met :
1. Users at every second (current _t ime) t raverse all of t heir UT XO, subst it ut ing int o
t he above f ormula t o see if it can sat isf y t he inequalit y condit ion. If it is sat isfied,
record t he corresponding UT XO in t he block and release t he block. (see Point 4)
2. St ake_modifier is t he value af t er hashing some of t he fields in t he previous block.
T his is added t o prevent users f rom knowing in advance when t hey have t he right t o
mine.
3. Difficult y will be dynamically adjust ed according t o t he recent block out put t ime t o
ensure a st able block generat ion t ime int erval.
4. Since we only need t o complet e t he hash calculat ion equal t o t he number of UT XOs
per second, t he required comput ing power is lower.
5. From t he inequalit y equat ion, we can see t hat t he more UT XOs are held, t he great er
t he amount of t okens in UT XO (coin(UT XO)), t he longer UT XO holds (age (UT XO)
or t he age of t he coin), and t he easier t he inequalit y is t o solve, t he easier it is t o
mine.
6. Generat e block reward set t ings f or coin(UT XO) * age(UT XO). T hat is, t he larger t he
UT XO amount and t he longer t he holding t ime, t he higher t he reward.
7. In order t o record eligible UT XO int o blocks and be compat ible wit h t he original PoW
mode, Peercoin designed t he logic of coinst ake: Keep t he original first t ransact ion
as coinbase, but t he required input quant it y must be equal t o 1, and t he input
"prev.out " field must be set t o a null value, and t he out put quant it y must be great er
t han or equal t o 1. If t he second t ransact ion needs t o be coinst ake, t his requires
t he input quant it y t o be great er t han or equal t o 1, and t he first input is UT XO t hat
meet s t he condit ion, t he out put quant it y is great er t han or equal t o 2, and t he first
out put must be blanked, and t he second out put is block reward.
hash(st ake_modifier, current _t ime, UT XO) < coin(UT XO) * age(UT XO) * difficult y
Due t o t he int roduct ion of t ime f act or, t here is t he possibilit y of a Coin Age
Accumulat ion At t ack. T hat is, t he node is shut down, and when t he age (UT XO) is large
enough, t he node mining is st art ed, t hereby saving power, which causes t he problems
t hat t he number of online nodes is t oo small and t he syst em is f ragile. Of course, it is
fine t o set an age limit , but doing so also loses t he meaning of age as a moderat or. In
summary, t he hybrid consensus Pillars will not consider t he scheme of coin age.
Potential Attacks
Bribe At t ack
T he bribe at t ack process is as f ollows:
1. T he at t acker buys a good or service.
2. Merchant s wait f or t he net work t o confirm t he deal.
3. At t his point , t he at t acker begins t o claim f or t he first t ime in t he net work and
rewards t he current longest chain t hat does not cont ain t his t ransact ion.
4. When t he main chain is long enough, t he at t acker begins t o give out more rewards
t o miners who mine t he chain t hat cont ains t he t ransact ion.
5. T he at t acker gives up t he reward af t er six confirmat ions are reached.
6. When t he goods arrive, t he at t acker gives up t he chain he select ed originally.
T heref ore, as long as t he cost of t he bribery at t ack is less t han t he cost of t he goods
or services, t he at t ack is successf ul. In cont rast , bribery at t acks in t he PoW mechanism
require t he bribery of most miners, so t he cost is ext remely high and difficult t o realize.
Alt hough pure PoS has many problems, PoS has bet t er flexibilit y t han PoW, so we will
consider solving t hese problems in t he Met averse hybrid consensus Pillars.
When Q is great er t han or equal t o W=30, t he at t acker's Q chain is not accept ed;
When Q is less t han or equal t o W=30, t he at t acker's Q chain is st ill in a st at e where
t he securit y confirmat ion number is not reached;
When a PoS miner init iat es a Not hing-at -St ake at t ack at block B, it is det ermined by
t he subsequent PoW block C. T he PoS miner cannot maint ain a longer f ork chain.
Ent ering t he t hree-way hybrid consensus st age, t he DPoS block f ollows t he PoW block.
T he DPoS block prevent s at t acks on A' and C'; on t he ot her hand, if DPoS t ries t o
at t ack it will be reject ed by PoW and PoS blocks.
While no consensus algorit hm can current ly resist a Sybil at t ack, t he Avat ar MARS
score makes such an at t ack significant ly more cost ly, t hereby est ablishing a
crypt oeconomic incent ive st ruct ure t hat reduces t he probabilit y of Sybil at t acks t o
near-zero.
1. A malicious act or could t amper wit h t he PoW block version, causing consensus
f ailure due t o a PoW block f alsely modified t o have a PoS block st ruct ure. A valid
PoS block must meet t hese condit ions:
a. block version is block_version_pos;
b. the block's first entry must be a coinbase tx;
c. the block's second entry must be a coinstake tx;
d. first coinstake tx input entry must represent the miner's stake UTXO;
e. first coinstake tx output entry must be empty;
f. second coinstake tx output corresponds to first input entry UTXO;
g. the witness's private key pair corresponds to the blocksig contained in the block
header;
A PoS block can only be valid when t he above condit ions are met , prevent ing a PoS-
PoW conf usion at t ack.
2. A malicous act or could t amper wit h t he PoS block version, causing consensus
f ailure due t o a PoS block f alsely modified t o have PoW block st ruct ure. A valid
PoW block must meet t hese condit ions:
a. block version is block_version_pow;
b. the block's first entry must be a coinbase tx;
c. there are no coinstake transactions in the block;
A PoW block can only be valid when t he above condit ions are met , prevent ing a PoW-
PoS conf usion at t ack.
T he Met averse t ime-window is adjust ed t o 38 seconds. Since PoW and PoS t imest amp
validat ions are diff erent , t hus it is impossible t o cause consensus f ailure by t imest amp
conf usion at t ack. Wit h an average block t ime of 25 seconds, an at t acker could at most
produce 2 at t acking blocks; any more, and t he difficult y increases, slowing block
product ion and prevent ing a t imest amp at t ack.
T he t heoret ical limit f or t he P2P net work t o process all blocks is about 10 seconds.
Current ly Met averse block t imes are not suit able f or opt imizat ion t o around 15 seconds
wit hout f urt her improvement of t he P2P net work and wit hout uncle blocks.
T he first st age of MPC will support Light ning Net work, a second-layer net work which
depends on t he underlying blockchain f or securit y. By using real Bit coin-like t ransact ions
and using it s nat ive smart -cont ract script ing language, it is possible t o creat e a secure
net work of part icipant s wit h high t hroughput wit hout significant ly compromising
securit y. Since Met averse t ransact ion st ruct ure is similar t o t he Bit coin net work, it is
relat ively simple t o develop our own LN implement at ion.
In order f or MST t o share t he asset dist ribut ion f unct ionalit y of Met averse consensus
algorit hm, MPC Phase 1 support s PoW mining and PoS mint ing f or MST asset s. In t he
second phase of MPC, PoW, PoS, and DPoS mining and mint ing will be support ed.
All Met averse asset st ruct ures are UT XO-based, and we have made out put ext ensions
f or t he coinbase as f ollows:
1 {
2 "hash" : "bf4ca43a2c23c8d5b06",
3 "inputs" :
4 [
5 {
6 "previous_output" :
7 {
8 "hash" : "00000000000000000000",
9 "index" : 4294967295
10 },
11 "script" : "[ b71f0f ]",
12 "sequence" : 0
13 }
14 ],
15 "outputs" :
16 [
17 {
18 "address" : "Miner's Address'",
19 "attachment" : // nomarl coinbase
20 {
21 "type" : "etp"
22 },
23 "index" : 0,
24 "value" : 94338400
25 },
26 {
27 "address" : "Miner's Address",
28 "attachment" : // new MST output for coinbase
29 {
30 "quantity" : 300000000,
31 "symbol" : "TEST001.MINING",
32 "type" : "asset-transfer"
33 },
34 "index" : 1,
35 "value" : 0
36 }
37 ],
38 }
Ref er t o MIP-15
Avat ars inherit OIDC when a Met averse wallet is opened, providing decent ralized ident it y
services and allowing t he ext ension of Avat ars t o t radit ional int ernet applicat ions.
T here is a cont inual challenge f acing blockchain applicat ions, finding t he balance
bet ween T PS and decent ralizat ion. When t he T PS is upgraded, t his hurt s
decent ralizat ion, so we consider t he double-chain archit ect ure. T he current Met averse
inf rast ruct ure act s as a f oundat ion t hat has decent ralizat ion capabilit y. T he second
chain provides high T PS t ransmission capacit y and can be synchronized wit h t he DPoS
on t he main chain. T he problem wit h t he second chain is how t o adapt t o t he exist ing
syst em.
T o be writ t en...
Binary Port
T here are many archit ect ural pat t erns, and we only discuss single-point applicat ions,
layered archit ect ure pat t erns, event -driven archit ect ural pat t erns, and micro-service
archit ect ure pat t erns. We will discuss where MBaaS should be in t hese archit ect ural
pat t erns.
First of all, MBaaS is a collect ion of services. T he represent at ion in t he syst em is a t ype
of service process, which is usually generat ed by t he wallet program.
Current ly we can operat e t wo modes:
1.) Wallet -segregat ed mode: Separat e f unct ions f rom t he wallet program int o a mult i-
process mode, wit h each process providing a light weight MBaaS;
2.) Wallet program unified mode: T he wallet program provides MBaaS, but can f orm a
mast er-slave relat ionship and make an int ernal dist ribut ed net work inst ead of
connect ing t o t he public net work. T he unified model put s higher demands on t he
opt imizat ion and st abilit y of t he wallet .
Unified Mode
T he wallet program at least provides int ernal high-speed synchronizat ion, and t he
int ernal nodes change f rom final consist ency t o st rong consist ency, which requires t hat
int ernal nodes can achieve st rong consist ency when t he blockchain f orks.
Separat ion mode and unified mode are not absolut e, and t here may be a mixt ure in t he
act ual applicat ion. We will now discuss t he archit ect ural model.
Monolith Applications
Single-point applicat ions are divided int o client -side single-point applicat ions and
server-side single-point applicat ions. An example of t he server-side single-point
applicat ion is WordPress. If we want WordPress t o support MST , t he quickest way is
t o build t he Met averse wallet on t he t he WordPress backend, and t hen modif y t he
backend code t o call t he MST relat ed API. T he final int erf ace will display t he MST
t oken. T his sit uat ion is suit able f or a unified mode and rapid deployment . Single-point
applicat ions are more commonly used in t he Microkernel Archit ect ure mode. For
example, t he Met averse Avat ar is embedded in t he Eclipse IDE, which requires t he
Met averse light wallet t o be plugged int o Eclipse as a plugin. T his sit uat ion is suit able
f or segregat ed mode, such as a light wallet .
Layered Architecture
Layered archit ect ure is suit able f or bot h segregat ed and unified modes depending on
t he scope of t he layer st ruct ure. Segregat ed mode is clearly more suit able f or large-
scale layers such as a Service-orient ed archit ect ure (SOA). In segregat ed mode, MBaaS
can be placed on t he business layer as a st andard compenent as t he wallet API only
needs t o be compat ible wit h ot her modules on t he same layer. St ruct ural blockchain
st orage may be required f or a persist ent dat a layer, ot herwise t he wallet it self can
direct ly replace t he block st orage f unct ion. In unified mode, MBaaS is suit ed f or small-
scale applicat ions where it can direct ly ref er t o t he server node.
Event-driven Architecture
Event -driven archit ect ure is suit able f or segregat ed models. T his model f ocuses on t he
dist ribut ion and processing of event s. Looking at t he logic of blockchain, we can see
t hat t he blockchain is based on t ransact ions and a t ransact ion it self is an event .
In Mediat or mode, t he abilit y t o resolve t ransact ions is needed t o parse and redist ribut e
t ransact ion t ypes and dat a. For account st at us model, it must also read account
st at us; in Broker mode, each Processor can parse and det ermine t he t ransact ion
wit hout involving changes made by t he Broker.
T he above process is input as an event , and when t ransact ion out put is required, we
can t hink of t he wallet as a processor and only deal wit h t he business relat ed t o t he
blockchain, but here we may encount er t he problem t hat t he processor evolves int o a
cent ral processor. Because t he ult imat e goal of any core business flow is payment , t he
wallet processor will become a collect ion of aut hent icat ion, signat ure, and broadcast
t ransact ions, and will encount er significant perf ormance bot t lenecks.
T heref ore, t he net work module and t ransact ion verificat ion module in segregat ed mode
can be horizont ally ext ended. T o accomplish t his Met averse should provide a complet e
SDK t o support event dist ribut ion and processing.
Microservice Architecture
Microservice archit ect ure is suit able f or bot h unified and segregat ed models. In a unified
model t he wallet becomes a microservice component as long as t he wallet f unct ions are
sufficient ly cohesive. For example, a wallet can perf orm payment in component A and
perf orm t ransact ion validat ion in component B. T his requires t he wallet f unct ions t o
adhere t o t he microservices archit ect ure and t o provide a robust query-verif y API. On
t he ot her hand, t he archit ect ural concept of microservices fit s very well wit h segregat ed
models, t hus it is not difficult t o st andardize Met averse microservice component s.
Smart Contracts(@Galaxy)
References
《PoW, PoS, & Hybrid prot ocols: A Mat t er of Complexit y?》ht t ps://arxiv.org/pdf /1
805.08674.pdf
《2-hop Blockchain: Combining Proof -of -Work and Proof -of -St ake Securely》ht t
ps://eprint .iacr.org/2016/716.pdf
ht t ps://git hub.com/mvs-org/mips/blob/mast er/mips/mip-2.md
ht t ps://git hub.com/Nevacoin/nevacoin
ht t ps://git hub.com/dashpay/dash/issues/2268
ht t ps://git hub.com/mvs-org/mips/blob/mast er/mips/mip-16.md
ht t p://newmet averse.org/whit e-paper/Met averse-whit epaper-v3.0-EN.pdf