Research Square (Research Square),
Год журнала:
2023,
Номер
unknown
Опубликована: Март 17, 2023
Abstract
Infrastructure
service
model
provides
different
kinds
of
virtual
computing
resources
such
as
networking,
storage
service,
and
hardware
per
user
demands.
Host
load
prediction
is
an
important
element
in
cloud
for
improvement
the
resource
allocation
systems.
Hosting
initialization
issues
still
exist
due
to
this
problem
takes
several
minutes
delay
response
process.
To
solve
issue
techniques
are
used
proper
data
center
dynamically
scale
order
maintaining
a
high
quality
services.
Therefore
paper,
we
propose
hybrid
convolutional
neural
network
long
with
short-term
memory
host
prediction.
In
proposed
model,
vector
auto
regression
method
firstly
input
analysis
which
filters
linear
interdependencies
among
multivariate
data.
Then
enduring
computed
entered
into
layer
that
extracts
complex
features
each
central
processing
unit
machine
usage
components
after
suitable
modeling
temporal
information
irregular
trends
time
series
components.
all
process,
main
contribution
scaled
polynomial
constant
activation
function
most
kind
model.
Due
higher
inconsistency
center,
accurate
For
reason
paper
two
real-world
traces
were
evaluate
performance.
One
trace
Google
while
other
traditional
distributed
system.
The
experiment
results
show
our
achieves
state-of-the-art
performance
accuracy
both
datasets
compared
ARIMA-LSTM,
VAR-GRU,
VAR-MLP,
CNN
models.
IEEE Communications Surveys & Tutorials,
Год журнала:
2023,
Номер
25(3), С. 1991 - 2020
Опубликована: Янв. 1, 2023
Future
communication
networks
are
envisioned
to
satisfy
increasingly
granular
and
dynamic
requirements
accommodate
the
application
user
demands.
Indeed,
novel
immersive
mission-critical
services
necessitate
increased
computing
network
resources,
reduced
latency,
guaranteed
reliability.
Thus,
efficient
adaptive
resource
management
schemes
required
provide
maintain
sufficient
levels
of
Quality
Experience
(QoE)
during
service
life-cycle.
Service
migration
is
considered
a
key
enabler
orchestration.
moving
on
demand
an
mechanism
for
mobility
support,
load
balancing
in
case
fluctuations
demands,
hardware
failure
mitigation.
However,
requires
planning,
as
multiple
parameters
must
be
optimized
reduce
disruption
minimum.
Recent
breakthroughs
computational
capabilities
allowed
emergence
Machine
Learning
tool
decision
making
that
expected
enable
seamless
automation
by
predicting
events
learning
optimal
policies.
This
paper
surveys
contributions
applying
(ML)
methods
optimize
migration,
providing
detailed
literature
review
recent
advances
field
establishing
classification
current
research
efforts
with
analysis
their
strengths
limitations.
Finally,
provides
insights
main
directions
future
research.
IEEE Access,
Год журнала:
2023,
Номер
11, С. 86739 - 86753
Опубликована: Янв. 1, 2023
The
energy
used
by
cloud
data
centers
(CDCs)
to
support
large
volumes
of
storage
and
computation
is
dramatically
increasing
as
the
scope
services
continues
expand.
This
puts
a
greater
burden
on
environment
results
in
higher
expenses
for
providers.
Virtualization
migration
consolidation
have
been
widely
current
CDCs
achieve
ser-vice
reduce
consumption
(EC).
study
divides
fundamental
tasks
virtual
machine
(VM)
into
three
portions:
determining
timing,
choosing
VMs
migrate
out,
selecting
destination
hosts.
An
EC
levels-based
adaptive
dynamic
threshold
method
timing
was
proposed,
well
correlation
utilization-based
strategy
out
an
improved
EC-aware
best-fit
algorithm
pro-posed
algorithms
were
evaluated
using
CloudSim
toolbox,
real
VM
workload
traces
from
PlanetLab
experimental
data.
According
experiments,
proposed
EC,
service
level
agreement
violation
(SLAV),
number
migrations
average
15.49%,
7.85%,
83.32%
comparison
related
state-of-the-art
methods
benchmark
algorithms.
suggests
that
outperform
other
techniques
migration,
even
when
necessitates
significant
or
amount
host
resources,
improve
quality
while
optimizing
consumption.
However,
experiments
conducted
simulation
platform,
which
has
some
drawbacks,
leading
varying
slightly
actual
environment.
Mathematics,
Год журнала:
2024,
Номер
12(3), С. 468 - 468
Опубликована: Фев. 1, 2024
The
advancement
of
cloud
computing
technologies
has
positioned
virtual
machine
(VM)
migration
as
a
critical
area
research,
essential
for
optimizing
resource
management,
bolstering
fault
tolerance,
and
ensuring
uninterrupted
service
delivery.
This
paper
offers
an
exhaustive
analysis
VM
processes
within
infrastructures,
examining
various
types,
server
load
assessment
methods,
selection
strategies,
ideal
timing,
target
determination
criteria.
We
introduce
queuing
theory-based
model
to
scrutinize
dynamics
between
servers
in
environment.
By
reinterpreting
resource-centric
mechanisms
into
task-processing
paradigm,
we
accommodate
the
stochastic
nature
demands,
characterized
by
random
task
arrivals
variable
processing
times.
is
specifically
tailored
scenarios
with
two
three
VMs.
Through
numerical
examples,
elucidate
several
performance
metrics:
blocking
probability,
average
tasks
processed
VMs,
managed
servers.
Additionally,
examine
influence
arrival
rates
duration
on
these
measures.
Acta Polytechnica Hungarica,
Год журнала:
2024,
Номер
21(6), С. 33 - 52
Опубликована: Янв. 1, 2024
Data
science
and
artificial
intelligence
are
emergently,
very
fast-evolving
fields,
being
applied
to
a
large
diversity
of
real-life
problem-solving.In
this
context,
some
methods
without
verifying
assumptions
that
must
be
met,
for
the
correct
applicability
necessary
model
fit.Such
mistakes
could
lead
misinterpretations
results.One
application
domains,
is
affected
in
sense,
healthcare,
where
have
dangerous
effects
on
human
health.Based
an
indepth
study
scientific
literature,
it
was
identified
bivariate
linear
regression
(BLR)
even
considered
simple,
one
sometimes
leads
confusion
application.With
mind,
paper
proposes
algorithmic
form
methodology
consists
assumptions,
passed
by
BLR,
so
should
pass
required
threshold
fit.Also,
presented
decision
calculus
correlation
coefficient
(BCC).There
other
considerations,
like
sample
sizes
two
variables
case
BCC
BLR.The
proposed
methodology,
herein,
will
useful
researchers,
since
BLR
frequently
research
diverse
industry
individually
or
combined
with
data
intelligence.
Energies,
Год журнала:
2024,
Номер
17(24), С. 6439 - 6439
Опубликована: Дек. 20, 2024
Energy
management
in
smart
cities
has
gained
particular
significance
the
context
of
climate
change
and
evolving
geopolitical
landscape.
It
become
a
key
element
sustainable
urban
development.
In
this
context,
energy
plays
central
role
facilitating
growth
cities.
The
aim
article
is
to
analyse
existing
scientific
research
related
cities,
identify
technological
trends,
highlight
prospective
directions
for
future
studies
field.
involves
literature
review
based
on
analysis
articles
from
Scopus
Web
Science
databases
evaluate
concerning
findings
suggest
that
should
focus
development
grids,
storage,
integration
renewable
sources,
as
well
innovative
technologies
(e.g.,
Internet
Things,
5G/6G,
artificial
intelligence,
blockchain,
digital
twins).
This
emphasises
can
enhance
efficiency
contributing
their
recommended
practical
policy
grids
cornerstone
adaptive
underpinned
by
regulations
encouraging
collaboration
between
operators
consumers.
Municipal
policies
prioritise
adoption
advanced
technologies,
such
IoT,
AI,
twins,
storage
systems,
improve
forecasting
resource
efficiency.
Investments
zero-emission
buildings,
renewable-powered
public
transport,
green
infrastructure
are
essential
enhancing
reducing
emissions.
Furthermore,
community
engagement
awareness
campaigns
form
an
integral
part
promoting
practices
aligned
with
broader
objectives.
Transactions on Emerging Telecommunications Technologies,
Год журнала:
2022,
Номер
34(1)
Опубликована: Окт. 3, 2022
Abstract
Workload
prediction
is
the
necessary
factor
in
cloud
data
center
for
maintaining
elasticity
and
scalability
of
resources.
However,
accuracy
workload
very
low,
because
redundancy,
noise,
low
center.
Therefore,
this
article,
a
tree
hierarchical
deep
convolutional
neural
network
(T‐CNN)
optimized
with
sheep
flock
optimization
algorithm
based
work
load
proposed
sustainable
centers.
Initially,
historical
from
preprocessed
using
kernel
correlation
method.
The
T‐CNN
approach
used
dynamic
environment.
weight
parameters
model
are
by
algorithm.
COSCO2
method
has
accurately
predicts
upcoming
reduces
extravagant
power
consumption
at
evaluated
utilizing
two
benchmark
datasets:
(i)
NASA,
(ii)
Saskatchewan
HTTP
traces.
simulation
implemented
java
tool
calculated.
From
simulation,
attains
20.64%,
32.95%,
12.05%,
32.65%,
26.54%
high
accuracy,
27.4%,
26%,
23.7%,
34.7%,
36.5%
lower
energy
validating
NASA
dataset,
similarly
20.75%,
19.06%,
29.09%,
23.8%,
20.5%
20.84%,
18.03%,
28.64%,
30.72%,
33.74%
traces
dataset
than
existing
approaches,
like
auto
adaptive
differential
evolution
BiPhase
learning‐based
network,
error
preventive
score
time
series
forecasting
models,
methods
prediction,
self‐directed
