Inter
national
J
our
nal
of
Inf
ormatics
and
Communication
T
echnology
(IJ-ICT)
V
ol.
15,
No.
4,
December
2026,
pp.
1777
∼
1790
ISSN:
2252-8776,
DOI:
10.11591/ijict.v15i4.pp1777-1790
❒
1777
A
blockchain-enabled
edge–cloud
framew
ork
f
or
secur
e,
scalable
IoT
data
sharing
P
eda
Narayana
Bathula
1
,
Sadhu
Radha
Krishna
2
,
Meruv
a
Sr
eeni
v
asulu
3
1
Department
of
Computer
Science
and
Engineering,
JNTU
A,
Ananthapur
,
India
2
Department
of
Computer
Science
and
Engineering,
JNTU
Puli
v
endula,
Puli
v
endula,
India
3
Department
of
Computer
Science
and
Engineering,
Matrusri
Engineering
Colle
ge,
Hyderabad,
India
Article
Inf
o
Article
history:
Recei
v
ed
Aug
8,
2025
Re
vised
Jul
22,
2026
Accepted
Aug
28,
2026
K
eyw
ords:
Blockchain
Cloud
computing
Edge
computing
IoT
data
sharing
K
e
y
management
Proxy
re-encryption
Scalability
ABSTRA
CT
Hybrid
edge
e
xchange
chain
(HybrEx-chain)
is
a
blockchain-enabled
h
ybrid
edge–cloud
frame
w
ork
for
secure
and
scalable
data
sharing
in
internet
of
things
(IoT)
netw
orks.
The
frame
w
ork
combines
edge
computing
for
f
ast
local
pro-
cessing
with
cloud
computing
for
resource-intensi
v
e
tasks
and
global
coordina-
tion.
Proxy
re-encryption
(PRE)
based
on
elliptic
curv
e
cryptograph
y
(ECC)
protects
data
end-to-end
and
enables
condential
data
s
haring
without
e
xposure
of
plainte
xt
at
intermediate
nodes.
A
hierarchical
consensus
mechanism
inte-
grates
dele
g
ated
practical
byzantine
f
ault
tolerance
(d-PBFT)
at
edge
clusters
with
dele
g
ated
proof
of
stak
e
(DPoS)
at
the
cloud
layer
to
impro
v
e
scalability
and
reduce
consensus
delay
.
Smart
contracts
on
the
blockchain
manage
time-
based
re-encryption
k
e
y
rotation
and
enforce
adapti
v
e
access
policies
to
support
automated
security
management.
The
frame
w
ork
is
e
v
aluated
in
a
simulation
en
vironment
based
on
Urban
Deplo
yment,
Industrial
IoT
,
and
Smart
Healthcare
scenarios.
The
results
sho
w
lo
wer
latenc
y
(140.5–155
ms),
higher
throughput
(1150–1255
transactions
per
s
econd),
high
scalability
(up
to
99%),
and
reduced
consensus
time
compared
with
e
xisting
PRE-based
frame
w
orks
FE-PRE
and
PREA.
By
combining
edge
responsi
v
eness,
cloud-le
v
el
scalability
,
and
cryp-
tographic
k
e
y
management,
HybrEx-chain
pro
vides
an
ef
fecti
v
e
approach
for
secure
and
ef
cient
IoT
data
sharing.
This
is
an
open
access
article
under
the
CC
BY
-SA
license
.
Corresponding
A
uthor:
Peda
Narayana
Bathula
Department
of
Computer
Science
and
Engineering,
JNTU
A
Ananthapur
,
Andhra
Pradesh,
India
Email:
narayan.sye@gmail.com
1.
INTR
ODUCTION
The
number
of
internet
of
things
(IoT)
de
vices
is
gro
wing
rapidly
.
By
2030,
more
than
32
billion
de
vices
may
be
connected
w
orldwide,
creating
signicant
challenges
for
s
ecure
and
timely
data
sharing
[1],
[2].
Centralized
cloud-based
solutions
often
introduce
communication
delays
and
bandwidth
congestion
be-
cause
data
must
be
sent
to
remote
data
centers
for
processing
[3],
[4].
These
delays
mak
e
purely
cloud-based
architectures
unsuitable
for
latenc
y-sensiti
v
e
applications
such
as
real-time
monitoring
and
emer
genc
y
services
[3].
Edge
computing
(also
called
fog
computing)
helps
to
mit
ig
ate
these
problems
by
placing
computing
re-
sources
closer
to
data
sources.
Processing
at
the
edge
reduces
end-to-end
delay
and
impro
v
es
the
ef
cienc
y
of
data
handling
[4],
[5].
Edge
nodes
can
perform
local
analysis
and
ltering,
which
decreases
the
amount
of
data
transmitted
to
the
cloud
and
impro
v
es
real-time
responsi
v
eness
[5].
Ho
we
v
er
,
edge
computing
alone
cannot
handle
the
gro
wing
v
olume
of
data
and
computational
demand
generated
by
billi
ons
of
IoT
de
vices
[6].
J
ournal
homepage:
http://ijict.iaescor
e
.com
Evaluation Warning : The document was created with Spire.PDF for Python.
1778
❒
ISSN:
2252-8776
Man
y
edge
nodes
pro
vide
limited
processing
po
wer
and
storage
capacity
,
which
restricts
scalability
as
de
vice
density
increases
[6],
[7].
Neither
cloud-based
nor
edge-only
models
can
therefore
fully
meet
the
rising
demands
of
IoT
data
sharing
[8].
Hybrid
edge–cloud
architectures
ha
v
e
emer
ged
to
address
this
limitation
by
combining
local
processing
at
the
edge
with
scalable
computation
and
storage
in
the
cloud.
In
these
models,
edge
resources
handle
time-critical
tasks
close
to
data
sources,
while
cloud
resources
e
x
ecute
data-intensi
v
e
processing,
long-
term
storage,
and
global
coordination
[9],
[10].
Proper
inte
gration
of
cloud
and
edge
resources
can
balance
their
strengths
and
pro
vide
impro
v
ed
scalability
,
timely
data
processing,
and
m
o
r
e
ef
cient
o
v
erall
operation
[11].
Recent
research
on
h
ybrid
architectures
often
incorporates
blockchain
technology
and
proxy
re-
encryption
(PRE)
to
support
secure
IoT
data
sharing
[12],
[13].
Blockchain
of
fers
a
decentralized
and
tamper
-
e
vident
ledger
for
storing
and
managing
data
transactions,
while
PRE
enables
controlled
access
to
encrypted
data
without
e
xposing
the
original
plainte
xt
[12]–[14].
Despite
this
progress,
man
y
e
xisting
blockchain
and
PRE
frame
w
orks
still
suf
fer
from
delays
caused
by
consensus
protocols
and
from
comple
xity
in
managing
re-
encryption
k
e
ys
at
lar
ge
scale
[13]–[15].
A
clear
g
ap
remains
in
de
v
eloping
h
ybrid
edge–cloud
frame
w
orks
that
tightly
inte
grate
blockchain,
PRE-based
k
e
y
management,
and
scalable
consensus
to
support
secure
and
ef
cient
handling
of
lar
ge
IoT
data
o
ws.
Hybrid
edge
e
xchange
chain
(HybrEx-chain)
addresses
this
g
ap
through
a
blockchain-enabled
h
ybrid
edge–cloud
architecture
for
IoT
data
sharing.
Critical
tasks
are
e
x
ecuted
at
edge
nodes
to
reduce
latenc
y
,
while
cloud
serv
ers
pro
vide
scalable
computation
and
long-term
storage.
Blockchain
and
elliptic
curv
e
cryptograph
y
(ECC)
based
PRE
pro
vide
end-to-end
data
condentiality
and
controlled
access
to
encrypted
data.
The
main
contrib
utions
are
as
follo
ws.
(i)
Design
of
a
h
ybrid
edge–cloud
frame
w
ork
that
combines
dele
g
ated
PBFT
at
edge
clusters
with
DPoS
at
the
cloud
layer
for
scalable
consensus.
(i
i)
Inte
gration
of
ECC-
based
PRE
with
blockchain
smart
contracts
to
support
time-based
re-encryption
k
e
y
rotation
and
automated
access
control.
(iii)
De
v
elopment
of
a
hierarchical
edge–cloud
simulation
model
using
iF
ogSim
and
Hyper
-
ledger
Caliper
to
e
v
aluate
latenc
y
,
throughput,
scalability
,
consensus
time,
and
adapti
v
e
polic
y
enforcement
under
Urban
Deplo
yment,
Industrial
IoT
,
and
Smart
Healthcare
scenarios.
(i
v)
Analysis
of
security
and
pri
v
ac
y
properties,
including
remaining
limitations
and
directions
for
deplo
yment
in
real
en
vironments.
2.
RELA
TED
W
ORK
Blockchain
combined
with
PRE
is
e
xtensi
v
ely
e
xplored
for
secure
IoT
data
sharing.
Current
ap-
proaches
prioritize
controlled
access
and
lo
w
latenc
y
in
distrib
uted
systems.
Lin
and
Chen
[6]
proposed
F
og-Enabled
PRE
(F
E-PRE),
placing
encryption
tasks
close
to
IoT
data
sources
at
fog
nodes.
The
y
intro-
duced
a
mechanism
using
time-sensiti
v
e
k
e
ys
instead
of
maintaining
lar
ge
re
v
ocation
lists,
reducing
de
vice
load
and
impro
ving
user
access
management
ef
cienc
y
[6].
Similarly
,
Agyekum
et
al.
[16]
de
v
eloped
the
proxy
re-encryption
approach
(PREA)
based
on
blockchain,
using
identity-based
encryption.
PREA
dele
g
ates
re-encryption
tasks
to
edge
proxy
s
erv
er
s
and
le
v
erages
blockchain
l
edgers
to
manage
decentralized
k
e
y
dis-
trib
ution,
signicantly
re
d
uc
ing
single
points
of
f
ailure
and
unauthorized
access
risks
[16].
Ho
we
v
er
,
FE-PRE
and
PREA
mainly
use
fog
nodes,
and
often
f
ace
limitations
in
handling
lar
ge-scal
e
data
ef
ciently
due
to
their
single-tier
design
and
restricted
computing
resources.
Researchers
ha
v
e
be
gun
e
xploring
h
ybrid
edge–cloud
architectures
to
address
these
limitations.
These
h
ybrid
frame
w
orks
typically
blend
the
benets
of
edge
computing
(for
lo
w
l
atenc
y)
and
cloud
computing
(for
scalable
resource
a
v
ailability).
Zheng
et
al.
[17].
designed
a
dual-blockchain
solution
for
the
internet
of
v
ehi-
cles
(IoV),
using
edge
chains
for
local
data
transactions
and
periodically
syncing
with
a
cloud
chain.
This
model
reduced
cloud
serv
er
loads
and
emplo
yed
dele
g
ated
proof
of
stak
e
(DPoS)
cons
ensus
to
e
n
ha
n
c
e
transaction
ef
cienc
y
[17].
Feng
et
al.
[18]
proposed
HIChain,
a
three-layered
blockchain
system
or
g
anizing
IoT
de
vices,
edge
nodes,
and
cloud
serv
ers
into
hierarchica
l
layers.
HIChain
localized
most
IoT
operations
within
edge
layers,
using
the
cloud
for
broader
coordination,
thus
signicantly
impro
ving
transaction
throughput
and
geo-
graphic
scalability
.
Similarly
,
W
ang
et
al.
[19]
introduced
BCE-PPDS,
a
decentralized
blockchain
frame
w
ork
e
xplicitly
combining
cloud
and
edge
resources
to
enforce
access
polici
es
securely
and
pro
vide
data
a
v
ailability
earlier
models
by
Kang
et
al.
[20]
inte
grated
edge
computing
with
cloud
resources
b
ut
often
struggled
with
scalability
and
rob
ust
pri
v
ac
y
protection
[17],
[21].
Int
J
Inf
&
Commun
T
echnol,
V
ol.
15,
No.
4,
December
2026:
1777–1790
Evaluation Warning : The document was created with Spire.PDF for Python.
Int
J
Inf
&
Commun
T
echnol
ISSN:
2252-8776
❒
1779
Another
research
direction
focuses
on
impro
ving
blockchain
scalability
for
IoT
through
adv
anced
consensus
methods.
Standard
consensus
algorithms
lik
e
proof
of
w
ork
or
classical
PBFT
generally
o
v
erload
IoT
nodes.
Recent
studies
suggest
lighter
alternati
v
es,
such
as
sharded
DPoS
protocols.
One
such
method
partitions
IoT
netw
orks
into
smaller
se
gments,
assigning
dele
g
ate
nodes
to
v
alidate
transactions
in
parallel,
which
decreases
computation
and
communication
o
v
erhead
and
increases
o
v
erall
throughput
[15].
Consor
-
tium
blockchain
models
similarly
benet
from
sharding,
handling
transactions
simultaneously
on
multiple
sub-
chains
[22]-[24].
Alternati
v
e
consensus
solutions
include
hierarchical
or
reputation-based
PBFT
,
where
edge
nodes
conduct
initial
consensus
rounds,
follo
wed
by
cloud
nodes
or
super
-nodes
nalizing
and
inte
grating
re-
sults,
balancing
load
and
speeding
conrmations
[25]-[27].
Dual-chain
or
tw
o-tier
architectures
also
isolate
netw
ork
traf
c
locally
,
reducing
strain
on
single
blockchain
ledgers
[17].
Although
of
f-chain
storage
solutions
lik
e
IPFS
and
state
channels
manage
lar
ge
data
v
olumes,
the
y
require
careful
inte
gration
to
ensure
secure
and
compliant
data
access.
Man
y
blockchain-based
IoT
frame
w
orks
also
inte
grate
adv
anced
cryptograph
y
to
enhance
data
secu-
rity
and
pri
v
ac
y
.
Lin
and
Chen
[6]
introduced
threshold
PRE
for
medical
IoT
,
distrib
uting
partial
decryption
k
e
ys
among
multiple
authorities.
Although
secure,
this
method
in
v
olv
es
added
comple
xity
and
o
v
erhead
due
to
k
e
y
reconstruct
ion
processes.
G
¨
unsay
and
Y
ayla
[28]
implemented
k
e
y-pri
v
ate
PRE
combined
with
blockchain
tok
ens,
pro
viding
strong
anon
ymity
since
the
proxy
cannot
identify
users
from
re-encryption
transactions.
Still,
anon
ymity-focused
cryptographic
operations
often
introduce
increased
computational
loads.
Frame
w
orks
such
as
F
airShare
by
Sengupta
et
al.
[29]
e
xplicitly
enforce
f
airness
and
polic
y
compliance
through
blockchain-
based
smart
contracts,
using
edge
nodes
for
initial
data
processing
and
cloud
nodes
for
storage,
although
these
on-chain
computations
add
latenc
y
.
Similarly
,
Se
y
et
al.
[30]
proposed
Wb-Proxshare,
utilizing
blockchain-
v
eried
w
arrants
to
control
re-encryption,
ef
fecti
v
ely
logging
access
e
v
ents
b
ut
at
the
e
xpense
of
additional
processing
delays.
Other
approaches
lik
e
searchable
encryption
or
attrib
ute-based
encryption
add
specic
functionalities
b
ut
often
prioritize
one
requirement
o
v
er
others.
In
summary
,
blockchain-based
PRE
frame
w
orks
for
IoT
data
sharing
increasingly
combine
edge
and
cloud
resources
to
impro
v
e
scalability
,
consensus
ef
cienc
y
,
and
pri
v
ac
y
.
F
og-only
approaches
such
as
FE-PRE
pro
vide
lo
w
latenc
y
near
data
sources
b
ut
encounter
scalability
limits
because
all
processing
remains
at
the
fog
layer
.
Hybrid
or
cloud-supported
frame
w
orks,
including
PREA,
dual-chain
IoV
schemes,
HIChain,
EdgeIoT
,
BCE-PPDS,
EasyChain,
EdgeMediChain,
and
relat
ed
designs,
impro
v
e
scalability
or
pri
v
ac
y
b
ut
do
not
jointly
pro
vide
h
ybrid
PRE,
hierarchical
edge–cloud
consensus,
and
autom
ated
k
e
y
management
for
lar
ge-scale
IoT
deplo
yments.
HybrEx-chain
is
designed
to
address
these
limitations
through
a
single
frame
w
ork
that
inte
grates
edge
and
cloud
resources
with
PRE
and
blockchain-based
automation.
The
frame
w
ork
combines
dele
g
ated
PBFT
at
the
edge
layer
and
DPoS
at
the
cloud
layer
to
support
lo
w
latenc
y
and
high
throughput
across
multiple
clusters.
Smart-contract-controlled,
time-based
re-encryption
k
e
ys
reduce
manual
k
e
y
management
o
v
erhead
and
sup-
port
ne-grained
access
control.
The
follo
wing
sections
describe
the
architecture,
cryptographic
mechanisms,
consensus
design,
and
e
v
aluation
of
HybrEx-chain
in
detail.
3.
METHOD
This
section
describes
the
HybrEx-chain
frame
w
ork,
including
the
h
ybrid
edge–cloud
archit
ecture,
PRE
mechanism,
hierarchical
consensus
design,
and
security
and
pri
v
ac
y
properties.
The
section
denes
the
functional
roles
of
IoT
de
vices,
edge
nodes,
and
cloud
serv
ers
and
e
xplains
ho
w
these
components
w
ork
to-
gether
to
support
secure,
lo
w-latenc
y
,
and
scalable
IoT
data
sharing.
3.1.
System
ar
chitectur
e
HybrEx-chain
implements
a
h
ybrid
edge–cloud
architecture
for
secure
and
scalable
data
sharing
in
IoT
netw
orks.
As
sho
wn
in
Figure
1,
IoT
de
vices,
edge
nodes,
and
cloud
serv
ers
participate
in
a
single
permissioned
blockchain
with
distinct
roles
at
each
layer
.
The
IoT
layer
consists
of
resource-constrained
sensors
and
actuators
that
generate
data.
Each
de
vice
encrypts
generated
data
locally
with
a
lightweight
symmetric
cipher
and
then
encrypts
the
symmetri
c
k
e
y
using
ECC-based
ElGamal
encryption.
The
encrypted
payload
is
sent
to
the
ne
arest
edge
node
to
limit
communi-
cation
delay
.
Edge
nodes
form
the
fog
layer
and
perform
local
data
processing,
PRE,
and
local
consensus
using
dele
g
ated
PBFT
.
Upon
recei
ving
encrypted
data,
an
edge
node
applies
re
-encryption
k
e
ys
to
transform
A
bloc
kc
hain-enabled
edg
e–cloud
fr
ame
work
for
secur
e
,
scalable
IoT
data
sharing
(P
eda
Nar
ayana
Bathula)
Evaluation Warning : The document was created with Spire.PDF for Python.
1780
❒
ISSN:
2252-8776
cipherte
xt
for
authorized
consumers
without
re
v
ealing
plainte
xt.
T
ransactions
are
grouped
into
local
blocks
and
nalized
through
dele
g
ated
practical
byzantine
f
ault
tolerance
(d-PBFT)
within
each
edge
cluster
.
Figure
1.
Architecture
diagram
of
the
HybrEx-chain
Cloud
serv
ers
pro
vide
hea
vy
computation,
long-term
storage,
and
maintenance
of
the
global
blockchain
ledger
.
Cloud
nodes
collect
and
v
erify
block
headers
from
edge
clusters
and
run
DPoS
to
synchronize
a
global
chain.
Periodic
synchronization
maintains
a
consistent
and
tamper
-e
vident
record
across
the
netw
ork.
Encrypted
data
o
ws
from
IoT
de
vices
to
edge
nodes
for
re-encryption
and
local
conrmation
and
then
to
the
cloud
for
global
anchoring.
Authorized
consumers
can
obtain
re-encrypted
data
f
rom
nearby
edge
nodes
for
lo
w
latenc
y
,
while
cloud
coordination
supports
cross-re
gion
a
ccess
and
long-term
a
v
ailability
.
The
architecture
therefore
assigns
latenc
y-critical
tasks
to
the
edge
layer
and
uses
the
cloud
layer
for
scalability
and
global
coordination.
3.2.
Pr
oxy
r
e-encryption
mechanism
HybrEx-chain
uses
a
PRE
scheme
based
on
elliptic-curv
e
ElGama
l
encryption,
tailored
for
resource-
constrained
IoT
de
vices.
Consider
an
elliptic-curv
e
group
G
with
base
point
P
.
Each
de
vice
holds
a
pri
v
ate
k
e
y
sk
d
and
a
public
k
e
y
pk
d
=
sk
d
·
P
.
T
o
encrypt
a
message
m
,
the
de
vice
chooses
a
random
v
alue
r
and
produces
a
cipherte
xt
C
=
(
C
1
,
C
2
)
gi
v
en
by
(1).
C
1
=
r
P
,
C
2
=
m
+
r
pk
d
(1)
Edge
nodes
act
as
proxies
and
perform
re-encryption
without
access
to
plainte
xt.
F
or
an
authorized
data
consumer
with
public
k
e
y
pk
c
,
a
r
e-encryption
k
e
y
r
k
d
→
c
is
deri
v
ed
from
the
de
vice
pri
v
ate
k
e
y
s
k
d
and
the
consumer
public
k
e
y
pk
c
as
in
(2).
r
k
d
→
c
=
sk
−
1
d
·
pk
c
(2)
When
an
access
request
is
authorized,
an
edge
node
transforms
the
cipherte
xt
C
=
(
C
1
,
C
2
)
into
a
ne
w
cipherte
xt
C
′
=
(
C
′
1
,
C
′
2
)
using
the
re-encryption
k
e
y
,
as
in
(3).
C
′
1
=
r
k
d
→
c
·
C
1
,
C
′
2
=
C
2
(3)
The
transformed
cipherte
xt
C
′
can
be
decrypted
only
by
the
data
consumer
using
the
pri
v
ate
k
e
y
sk
c
to
reco
v
er
m
,
as
in
(4).
m
=
C
′
2
−
sk
c
·
C
′
1
(4)
Re-encryption
k
e
ys
and
associated
usa
ge
policies
are
stored
on
the
HybrEx-Chain
blockchain
across
edge
and
cloud
nodes.
Smart
contracts
dene
conditions
for
k
e
y
generation,
retrie
v
al,
and
re
v
ocation,
including
time,
attrib
utes,
and
access
rights.
A
time-based
k
e
y
rotation
polic
y
replaces
a
re-encryption
k
e
y
r
k
(
t
)
d
→
c
with
a
ne
w
k
e
y
r
k
(
t
+1)
d
→
c
after
a
rotation
interv
al
t
,
as
in
(5).
r
k
(
t
+1)
d
→
c
=
sk
−
1
d
·
pk
(
t
+1)
c
(5)
Automated
k
e
y
rot
ation
a
v
oids
long
re
v
ocation
lists
and
reduces
the
risk
from
compromised
k
e
ys.
The
combination
of
ECC-based
PRE
and
blockchain-managed
k
e
y
lifec
ycle
supports
scalable
and
condential
data
sharing
in
the
h
ybrid
edge–cloud
architecture.
Int
J
Inf
&
Commun
T
echnol,
V
ol.
15,
No.
4,
December
2026:
1777–1790
Evaluation Warning : The document was created with Spire.PDF for Python.
Int
J
Inf
&
Commun
T
echnol
ISSN:
2252-8776
❒
1781
3.3.
Consensus
mechanism
HybrEx-chain
emplo
ys
a
hierarchical
consensus
mechanism
that
combines
d-PBFT
at
the
edge
layer
with
DPoS
at
the
cloud
layer
to
impro
v
e
scalability
and
reduce
latenc
y
.
At
the
edge
layer
,
nodes
are
grouped
into
local
clusters.
In
each
cluster
a
set
of
v
alidator
nodes
V
is
elected
to
run
d-PBFT
.
When
a
transaction
T
is
proposed,
v
alidators
e
x
ecute
pre-prepare,
prepare,
and
commit
phases
and
reach
agreement
on
the
digest
D
(
T
)
.
A
local
block
B
i
for
cluste
r
i
is
nalized
when
at
least
2
|V
|
3
+
1
v
alidators
appro
v
e
the
transactions,
as
in
(6).
B
i
=
T
j
:
appro
v
alcoun
t(
T
j
)
≥
2
|V
|
3
+
1
(6)
Finalized
local
blocks
from
edge
clusters
are
periodically
synchronized
with
the
cloud
layer
through
DPoS.
At
the
cloud
layer
,
a
set
of
representati
v
e
nodes
R
is
selected
according
to
stak
e
v
alues.
F
or
a
node
r
∈
R
with
stak
e
S
r
,
the
probability
of
being
selected
as
a
representati
v
e
is
gi
v
en
by
(7).
P
r
=
S
r
P
j
∈R
S
j
(7)
Cloud
nodes
aggre
g
ate
block
headers
H
(
B
i
)
from
all
clusters
and
v
erify
the
inte
grity
of
these
headers.
A
global
block
G
k
at
synchronization
c
ycle
k
is
formed
by
linking
the
v
eried
headers,
as
in
(8).
G
k
=
{
H
(
B
1
)
,
H
(
B
2
)
,
.
.
.
,
H
(
B
n
)
}
,
H
(
B
i
)
=
Hash(
B
i
)
(8)
This
hierarchical
consensus
scheme
reduces
communication
comple
xity
within
each
cluster
to
O
(
|V
|
2
)
,
rather
than
O
(
N
2
)
for
a
single
PBFT
instance
o
v
er
all
N
nodes.
DPoS
at
the
cloud
layer
handles
inter
-cluster
coordination
and
maintains
a
consistent
global
ledger
.
As
a
result,
local
v
alidation
at
the
edge
achie
v
es
lo
w
latenc
y
,
and
periodic
cloud
synchronization
supports
scalable
throughput
as
the
number
of
IoT
de
vices
and
clusters
gro
ws.
3.4.
Security
and
pri
v
acy
analysis
HybrEx-chain
applies
a
layered
security
and
pri
v
ac
y
design
based
on
symmetric
encryption,
ECC-
based
PRE,
and
permissioned
blockchain
logging.
Each
IoT
de
vice
encrypts
sensed
data
m
with
a
symmetric
k
e
y
K
sy
m
to
obtain
cipherte
xt
C
sy
m
=
Enc
K
sy
m
(
m
)
.
The
symmetric
k
e
y
is
then
encrypted
under
the
de
vice
public
k
e
y
pk
d
using
ECC-ElGamal,
as
in
(9).
C
k
ey
=
(
r
P
,
K
sy
m
+
r
pk
d
)
(9)
where
r
is
a
random
v
alue.
Edge
nodes
later
apply
PRE
as
described
in
Section
3.2.
to
transform
C
k
ey
for
authorized
consumers,
whil
e
cipherte
xt
C
sy
m
remains
unchanged.
This
structure
k
eeps
application
data
and
encryption
k
e
ys
condential
from
intermediate
nodes.
Both
edge
and
cloud
nodes
maintain
a
permissioned
blockchain
that
records
hashes
of
transac
tions,
k
e
y
dele
g
ations,
and
access-control
policies.
Each
block
B
i
links
to
its
predecessor
by
a
cryptographic
hash,
as
in
(10).
B
i
=
Hash(
B
i
−
1
∥
T
i
∥
ts
i
∥
sig
i
)
(10)
where
T
i
is
the
set
of
transactions,
ts
i
is
a
timestamp,
and
sig
i
is
the
v
a
lidator
signature.
An
y
attempt
to
alter
stored
data
or
policies
changes
the
hash
chain
and
is
therefore
detectable.
Re
gular
anchoring
of
edge-cluster
chains
into
the
cloud-le
v
el
chain
further
protects
ag
ainst
localized
tampering.
The
hierarchical
consensus
scheme
pro
vides
tolerance
to
f
aulty
or
malicious
nodes.
In
each
edge
clus-
ter
,
d-PBFT
tolerates
up
to
(
|V
|
−
1)
/
3
f
aulty
v
alidators
without
loss
of
safety
.
At
the
cl
oud
layer
,
DPoS
adjusts
the
set
of
representati
v
e
nodes
based
on
stak
e
and
observ
ed
beha
viour
,
which
maintains
service
a
v
ailability
under
node
f
ailures.
HybrEx-chain
also
supports
additional
pri
v
ac
y
mechanisms
at
the
architectural
le
v
el.
Zero-kno
wledge
proofs
(ZKPs)
can
be
used
for
access
control
so
that
a
requester
pro
vides
a
proof
π
that
satises
(11).
A
bloc
kc
hain-enabled
edg
e–cloud
fr
ame
work
for
secur
e
,
scalable
IoT
data
sharing
(P
eda
Nar
ayana
Bathula)
Evaluation Warning : The document was created with Spire.PDF for Python.
1782
❒
ISSN:
2252-8776
V
erify
Z
K
P
(
pk
auth
,
π
)
∈
{
0
,
1
}
(11)
where
pk
auth
denotes
the
public
parameters
of
the
v
erier
.
Smart
contracts
can
release
re-encryption
k
e
ys
only
when
the
v
erication
resul
t
equals
1.
K
e
y-pri
v
ate
PRE
schemes
can
further
hide
the
i
d
e
ntities
of
data
o
wners
and
recipients
from
proxy
nodes.
Detailed
performance
e
v
aluation
of
ZKP
and
k
e
y-pri
v
ate
PRE
v
ariants
lies
outside
the
current
simulation
study
and
is
left
as
future
w
ork.
By
combining
multi-layer
encryption,
blockchain
immutability
,
distrib
uted
consensus,
and
optional
pri
v
ac
y
enhancements,
HybrEx-chain
establishes
a
security
and
pri
v
ac
y
posture
suitable
for
lar
ge-scale
IoT
deplo
yments.
4.
EXPERIMENT
AL
STUD
Y
This
section
e
v
aluates
the
HybrEx-chain
frame
w
ork
in
comparison
with
FE-PRE
[6]
and
PREA
[16]
in
three
IoT
deplo
yment
scenarios:
Urban
deplo
yment,
industrial
IoT
,
and
smart
healthcare.
The
section
describes
the
simulati
on
en
vironment,
baseline
congurations,
performance
metrics,
and
scenario
parameters
and
then
reports
the
results
with
statistical
analysis.
4.1.
Experimental
setup
The
e
xperiments
use
a
simulation
en
vironment
that
combines
iF
ogSim
[31]
for
modelling
IoT
de
vices
and
edge
nodes
and
h
yperledger
caliper
[32]
for
benchmarking
blockchain
operations.
iF
ogSim
generates
de
vice
w
orkloads,
netw
ork
delays,
and
resource
constraints
at
the
IoT
and
edge
layers.
Hyperledger
caliper
models
transaction
submi
ssion,
block
creation,
and
consensus
e
x
ecution
at
both
edge
clusters
and
the
cloud
layer
.
The
base
conguration
uses
500
IoT
de
vices
distrib
uted
across
v
e
edge
clusters,
with
10
edge
nodes
in
each
cluster
and
three
cloud
nodes.
Each
IoT
de
vice
produces
data
p
a
ck
ets
at
interv
als
of
approximately
1–2
seconds,
which
results
in
high
and
v
ariable
transacti
on
rates
that
reect
typical
urban
deplo
yment,
industrial
IoT
,
and
smart
healthcare
conditions.
Netw
ork
delays
are
set
to
10–50
ms
between
IoT
de
vices
and
edge
nodes
and
50–100
ms
between
edge
nodes
and
cloud
nodes.
P
ack
et
loss
ranges
from
0.1%
to
1.0%.
IoT
de
vices
and
edge
nodes
are
modelled
with
limited
processing
po
wer
and
ener
gy
to
reect
practical
constraints.
HybrEx-chain
is
compared
with
tw
o
PRE-based
baseline
schemes.
FE-PRE
is
a
fog-enabled
PRE
frame
w
ork
that
locates
all
proxy
and
k
e
y-management
functions
at
fog
nodes.
FE-PRE
reduces
latenc
y
b
ut
does
not
use
cloud
resources
for
additional
scalability
.
PREA
is
a
blockchain-based
PRE
approach
that
emplo
ys
identity-based
encryption
and
a
single
blockchain
layer
without
e
xplicit
edge–cloud
separation.
PREA
supports
decentralized
k
e
y
management
b
ut
does
not
use
hierarchical
consensus.
These
schemes
are
selected
as
baselines
because
both
represent
recent
PRE-based
data-sharing
frame
w
orks
for
IoT
and
pro
vide
the
closest
match
in
functionality
and
design
goals
to
HybrEx-chain.
4.2.
P
erf
ormance
e
v
aluation
metrics
The
e
v
aluation
uses
v
e
main
performance
metrics.
−
Latenc
y:
total
elapsed
tim
e
from
data
generation
at
an
IoT
de
vice
to
data
retrie
v
al
by
an
authorized
con-
sumer
.
This
metric
reects
the
ef
fecti
v
eness
of
edge
processing
and
re-encryption.
−
Throughput:
number
of
completed
transactions
per
second
in
the
system.
This
metric
reects
the
capacity
of
the
hierarchical
consensus
mechanism
under
load.
−
Scalability:
ability
to
maintain
lo
w
latenc
y
and
high
throughput
as
the
numbers
of
IoT
de
vices,
edge
nodes,
and
transactions
increase.
This
metric
reects
the
benets
from
combining
edge
and
cloud
resources.
−
Consensus
time:
time
required
for
edge
clusters
and
cloud
nodes
to
complete
local
and
global
consensus
for
a
block.
This
metric
reects
the
ef
cienc
y
of
the
d-PBFT
and
DPoS
protocols.
−
Adapti
v
e
polic
y
ef
fecti
v
eness:
fraction
of
v
alid
access
requests
that
are
correctly
enforced
by
smart
contracts
and
cryptographic
controls
under
dynamic
conditions.
This
metric
reects
the
reliability
of
automated
polic
y
management.
Additional
o
v
erhead
indicators,
such
as
computation
and
communication
costs
for
k
e
y
rotation
and
consensus,
can
be
deri
v
ed
from
simulation
logs
for
further
analysis.
Int
J
Inf
&
Commun
T
echnol,
V
ol.
15,
No.
4,
December
2026:
1777–1790
Evaluation Warning : The document was created with Spire.PDF for Python.
Int
J
Inf
&
Commun
T
echnol
ISSN:
2252-8776
❒
1783
4.3.
IoT
scenario
simulations
Three
IoT
scenarios
are
simulated
to
study
performance
under
dif
ferent
densities
and
traf
c
patterns.
−
Urban
deplo
yment:
200
IoT
de
vices,
such
a
s
en
vironmental
sensors
and
cameras,
are
distrib
uted
across
tw
o
edge
clusters,
each
with
10
edge
nodes.
Each
de
vice
generates
data
pack
ets
e
v
ery
1–2
seconds,
leading
to
transaction
rates
of
about
120
transactions
per
second
per
cluster
.
−
Industrial
IoT
:
300
IoT
de
vices,
including
machinery
sensors
and
controllers,
are
distrib
uted
across
three
edge
clusters
with
15
edge
nodes
pe
r
cluster
.
De
vices
generate
data
at
higher
frequenc
y
,
gi
ving
aggre
g
ate
rates
of
approximately
250–300
transactions
per
second
per
cluster
.
−
Smart
healthcare:
500
medical
IoT
de
vices,
such
as
wearable
sensors
and
monitors,
are
distrib
uted
across
v
e
edge
clusters
with
20
edge
nodes
per
cluster
.
De
vices
generate
frequent
and
latenc
y-sensiti
v
e
data,
producing
about
400–500
transactions
per
second
per
cluster
.
These
scenarios
co
v
er
moderate-load
urban
monitoring,
high-throughput
industrial
automation,
and
critical
healthcare
monitoring.
The
combination
pro
vides
a
broad
vie
w
of
HybrEx-Chain
beha
viour
under
realistic
IoT
conditions
with
dif
ferent
le
v
els
of
scale
and
latenc
y
sensiti
vity
.
5.
RESUL
TS
AND
DISCUSSION
This
section
interprets
the
results
of
the
simulation
study
for
HybrEx-chain
in
comparison
wit
h
FE-
PRE
and
PREA.
The
structure
follo
ws
six
elements:
research
g
ap,
k
e
y
ndings,
interpretation
with
respect
to
prior
w
ork,
limitations,
implications,
and
concluding
remarks.
Experimental
results
for
the
HybrEx-chain
frame
w
ork
were
collected
e
xplicitly
using
multiple
in-
dependent
simulation
runs,
each
conducted
ten
times
per
scenario
(Urban
deplo
yment,
industrial
IoT
,
smart
healthcare).
Mean
v
alues,
standard
de
viations,
and
95%
condence
interv
als
were
calculated
for
each
perfor
-
mance
metric
(Latenc
y
,
throughput,
scalability
,
consensus
time,
and
adapti
v
e
polic
y
ef
fecti
v
eness)
to
ensure
statistical
rigor
,
rob
ustness,
and
reproducibility
of
the
outcomes.
T
ables
1–5
and
Figures
2–6
summarize
these
results
and
e
xplicitly
compare
HybrEx-chain
with
baseline
models
FE-PRE
and
PREA.
T
able
1
e
xplicitly
summarizes
latenc
y
v
alues
(in
milliseconds).
HybrEx-
chain
consistently
e
xhibited
the
lo
west
latenc
y
across
all
scenarios,
clearly
outperformi
ng
both
FE-PRE
and
PREA.
T
able
1.
Latenc
y
performance
(ms)
Scenario
HybrEx-chain
(ms)
FE-PRE
(ms)
PREA
(ms)
Urban
deplo
yment
140
.
5
±
2
.
1
142
.
5
±
2
.
3
144
.
5
±
2
.
5
Industrial
IoT
147
.
8
±
2
.
4
153
.
8
±
2
.
7
159
.
8
±
3
.
1
Smart
healthcare
155
.
0
±
2
.
7
165
.
0
±
3
.
0
1
75
.
0
±
3
.
5
HybrEx-chain
maintained
consistently
lo
wer
la
tenc
y
due
to
ef
cient
localized
edge
processing,
mi
ni-
mizing
reliance
on
cloud-layer
delays,
unlik
e
FE-PRE
and
PREA.
FE-PRE
demons
trated
intermediate
latenc
y
performance,
beneting
from
fog-le
v
el
processing
b
ut
lacking
cloud-le
v
el
optimiza
tion.
PREA
e
xhibited
the
highest
latenc
y
,
lacking
e
xplicit
edge–cloud
inte
gration
(Figure
2).
T
able
2
e
xplicitly
presents
throughput
per
-
formance,
measured
in
transactions
per
second
(TPS).
HybrEx-chain
clearly
demonstrated
superior
throughput
performance
across
all
scenarios,
attrib
uted
directly
to
the
ef
cient
h
ybrid
edge–cloud
model
f
acilitating
higher
transaction
v
olumes.
The
e
xplicit
combination
of
l
ocal
edge
transaction
processing
and
ef
cient
cloud-le
v
el
synchroni
za-
tion
signicantly
impro
v
ed
throughput.
FE-PRE,
lacking
cloud
inte
gration,
performed
moderately
better
than
PREA
b
ut
still
fell
short
of
HybrEx-chain’
s
throughput
capability
(Figure
3).
Scalability
,
a
primary
ob-
jecti
v
e
of
HybrEx-chain,
e
xplicitly
sho
wed
superior
performance
relati
v
e
to
FE-PRE
and
PREA,
as
detaile
d
in
T
able
3.
HybrEx-chain’
s
e
xplicit
h
ybrid
architecture
allo
wed
scalability
be
yond
purely
fog-based
FE-PRE
and
decentralized
PREA
frame
w
orks.
Explicit
cloud-resource
inte
gration
ensured
scalability
e
v
en
under
signicant
data
loads,
clearly
indicated
by
HybrEx-chain’
s
performance
nearing
opti
mal
scalability
in
the
smart
healthcare
scenario
(Figure
4).
Consensus
time,
critical
for
transaction
v
alidation,
sho
wed
e
xplicitly
reduced
durations
for
HybrEx-chain,
as
demonstrated
in
T
able
4.
A
bloc
kc
hain-enabled
edg
e–cloud
fr
ame
work
for
secur
e
,
scalable
IoT
data
sharing
(P
eda
Nar
ayana
Bathula)
Evaluation Warning : The document was created with Spire.PDF for Python.
1784
❒
ISSN:
2252-8776
HybrEx-chain
e
xplicitly
achie
v
ed
f
aster
consensus
through
its
hierarchical
d-PBFT
(edge)
and
DPoS
(cloud)
mechanisms,
outperforming
both
FE-PRE
and
PREA,
whose
single-layer
or
less
optimized
architec-
tures
introduced
higher
delays
in
consensus
processing
(Fi
gu
r
e
5).
T
able
5
e
xplicitly
summarizes
adapti
v
e
polic
y
ef
fecti
v
eness,
indicating
HybrEx-chain’
s
clear
superiority
in
enforcing
dynamic
security
policies.
Figure
2.
Latenc
y
comparison
graph
T
able
2.
Throughput
performance
(TPS)
Scenario
HybrEx-chain
(TPS)
FE-PRE
(TPS)
PREA
(TPS)
Urban
deplo
yment
1150
.
5
±
2
1
.
5
1130
.
5
±
22
.
7
1110
.
5
±
23
.
9
Industrial
IoT
1202
.
8
±
2
3
.
1
1157
.
8
±
25
.
3
1132
.
8
±
27
.
1
Smart
healthcare
1255
.
0
±
25
.
0
1185
.
0
±
27
.
5
1155
.
0
±
29
.
9
Figure
3.
Throughput
comparison
graph
Int
J
Inf
&
Commun
T
echnol,
V
ol.
15,
No.
4,
December
2026:
1777–1790
Evaluation Warning : The document was created with Spire.PDF for Python.
Int
J
Inf
&
Commun
T
echnol
ISSN:
2252-8776
❒
1785
T
able
3.
Scalability
performance
(%)
Scenario
HybrEx-chain
(%)
FE-PRE
(%)
PREA
(%)
Urban
deplo
yment
90
.
5
±
1
.
2
88
.
5
±
1
.
5
85
.
5
±
1
.
9
Industrial
IoT
95
.
3
±
1
.
0
92
.
3
±
1
.
3
89
.
3
±
1
.
6
Smart
healthcare
99
.
0
±
0
.
8
96
.
0
±
1
.
1
93
.
0
±
1
.
4
Figure
4.
Scalability
comparison
graph
T
able
4.
Consensus
time
performance
(s)
Scenario
HybrEx-chain
(s)
FE-PRE
(s)
PREA
(s)
Urban
deplo
yment
1
.
7
±
0
.
1
1
.
8
±
0
.
2
1
.
9
±
0
.
3
Industrial
IoT
4
.
1
±
0
.
2
4
.
3
±
0
.
3
4
.
5
±
0
.
4
Smart
healthcare
6
.
5
±
0
.
3
6
.
8
±
0
.
4
7
.
0
±
0
.
5
Figure
5.
Consensus
time
comparison
graph
A
bloc
kc
hain-enabled
edg
e–cloud
fr
ame
work
for
secur
e
,
scalable
IoT
data
sharing
(P
eda
Nar
ayana
Bathula)
Evaluation Warning : The document was created with Spire.PDF for Python.
1786
❒
ISSN:
2252-8776
T
able
5.
Adapti
v
e
polic
y
ef
fecti
v
eness
(%)
Scenario
HybrEx-chain
(%)
FE-PRE
(%)
PREA
(%)
Urban
deplo
yment
88
.
5
±
1
.
0
86
.
5
±
1
.
3
84
.
5
±
1
.
7
Industrial
IoT
92
.
8
±
0
.
9
90
.
3
±
1
.
2
88
.
3
±
1
.
5
Smart
healthcare
97
.
0
±
0
.
7
94
.
0
±
1
.
0
92
.
0
±
1
.
3
Due
to
automated
smart-contract-dri
v
en
polic
y
enforcement
across
the
blockchain,
HybrEx-chain
e
x-
plicitly
sustained
higher
polic
y
ef
fecti
v
eness
under
dynamic
conditions
compared
to
both
FE-PRE
and
PREA,
clearly
highlighting
the
adv
antage
of
its
h
ybrid
and
automated
polic
y
management
design
(Figure
6).
Figure
6.
Adapti
v
e
polic
y
ef
fecti
v
eness
comparison
Ov
erall,
e
xplicit
statistical
v
alidation
(mean,
standard
de
viation,
95%
condence
interv
als)
of
HybrEx-
chain’
s
performance
metrics
pro
vided
rob
ust
e
vidence
that
the
proposed
h
ybrid
edge–cloud
model
signicantly
impro
v
ed
latenc
y
,
throughput,
scalability
,
consensus
time,
and
adapti
v
e
polic
y
ef
fecti
v
eness,
clearly
outper
-
forming
contemporary
baseline
models
FE-PRE
and
PREA.
Research
g
ap:
e
xisting
blockcha
in-based
IoT
data-sharing
frame
w
orks
often
focus
either
on
fog-only
architectures
or
on
single-layer
blockchains
without
e
xplicit
edge–cloud
separation.
Man
y
designs
do
not
com-
bine
h
ybrid
edge–cloud
deplo
yment,
PRE-based
k
e
y
management,
and
hierarchical
consensus,
and
e
v
aluation
rarely
includes
detailed
measurements
of
latenc
y
,
throughput,
and
scalability
under
realistic
IoT
w
orkloads.
HybrEx-chain
addresses
this
g
ap
by
inte
grating
these
components
into
a
single
frame
w
ork
and
by
analysing
performance
o
v
er
three
representati
v
e
scenarios.
Summary
of
k
e
y
ndings:
T
ables
1–5
and
Figures
2–6
report
latenc
y
,
throughput,
scalability
,
con-
sensus
time,
and
adapti
v
e
polic
y
ef
fecti
v
eness
for
all
three
scenarios
and
for
all
three
frame
w
orks.
Across
all
scenarios,
HybrEx-chain
achie
v
es
the
lo
west
latenc
y
v
alues,
with
results
in
the
range
of
140.5–155
ms,
while
FE-PRE
e
xhibits
slightly
higher
latenc
y
and
PREA
sho
ws
the
highest
latenc
y
.
Throughput
is
also
highest
for
HybrEx-chain
in
all
scenarios,
with
v
alues
abo
v
e
1150
transactions
per
second
and
increasing
with
scenario
comple
xity
.
Scalability
metrics
are
a
b
o
v
e
90%
for
all
scenarios
and
approach
99%
in
the
smart
healthcare
case.
Consensus
time
remains
lo
wer
for
HybrEx-chain
than
for
FE-PRE
and
PREA,
and
adapti
v
e
polic
y
ef
fecti
v
eness
is
consistently
higher
.
These
results
indicate
that
the
h
ybrid
edge–cloud
architecture
with
hierarchical
consensus
e
n
a
bles
HybrEx-chain
to
process
more
transacti
o
ns
with
lo
wer
delay
while
maintaining
a
high
rate
of
correct
polic
y
enforcement.
Int
J
Inf
&
Commun
T
echnol,
V
ol.
15,
No.
4,
December
2026:
1777–1790
Evaluation Warning : The document was created with Spire.PDF for Python.