Você está na página 1de 7

A

Forecasting
Methodology
for
Workload
Forecasting in Cloud Systems
Forecasting is the process of making predictions of the future based on past
and present data and most commonly by analysis of trends. A commonplace
example might be estimation of some variable of interest at some specified future
date. Prediction is a similar, but more general term. Both might refer to formal
statistical methods employing time series, cross-sectional or longitudinal data, or
alternatively to less formal judgmental methods. Usage can differ between areas of
application: for example, in hydrology the terms "forecast" and "forecasting" are
sometimes reserved for estimates of values at certain specific future times, while
the term "prediction" is used for more general estimates, such as the number of
times floods will occur over a long period. Cloud Computing is an essential
paradigm of computing services based on the elasticity property, where available
resources are adapted efficiently to different workloads over time. In elastic
platforms, the forecasting component can be considered by far the most important
element and the differentiating factor when comparing such systems, being
workload forecasting one of the problems to solve if we want to achieve a truly
elastic system.. As there is no general methodology in the literature that addresses
this problem analytically and from a time series forecasting perspective (even less
in the cloud field), we propose a combination of these tools based on a state ofthe-art forecasting methodology which we have enhanced with some elements,
such as: a specific cost function, statistical tests, visual analysis, etc. The insights
obtained from this analysis are used to detect the asymmetrical nature of the
forecasting problem and to find the best forecasting from the viewpoint of the

current state of the art in time series forecasting. From an operational point of view
the most interesting forecast is a short-time horizon, so we focus on this. To show
the feasibility of this methodology, The results show that the analyzed series are
non-linear in nature, and that no seasonal patterns can be found. Moreover, on the
analyzed datasets, the penalty cost as usually included in the SLA can be reduced
down to a 30% on average.

SYSTEM ANALYSIS
EXISTING SYSTEM
While there are many solutions in the literature based
on existing forecasting tools, there is no general framework
or methodology that addresses the workload forecasting
problem analytically from a state-of-the-art time series
forecasting perspective; even less in the cloud field, where
we perceive a gap between forecasters and domain experts,
in the sense that well-structured state of- the-art forecasting
methodology is not commonly employed in this area.

For example, one of the main results of the M3


forecasting

competition

is

that

simple

methods

often

outperform more complex ones [6], and model performance


depends on the underlying data of a forecasting problem.
More recently, the sNN3 competition [7] yielded as a result
that none of the participating Machine Learning (ML)
methods was able to outperform the theta method, which is
equivalent to simple exponential smoothing with drift, which
is a relatively simple standard method in time series
forecasting
PROPOSED SYSTEM

In

conclusion,

the

application

of

the

proposed

methodology has led us to gain some useful insight


regarding CWF time series properties, i.e., no seasonality,
nor trend is found, and they are non-linear. This helps to
discard some models and focus on others. In addition, by
using

the

forecasting

models

developed

here,

significant reduction in under-provisioning and overprovision costs can be achieved control inherently. The
proposed protocols also protect the privacy of filenames.
PROPOSED SYSTEM ALGORITHMS
Time Series:

Time-series forecasting methods use historical information only to produce


estimates of future values. Time-series forecasting techniques assume the data's
past pattern will continue in the future and include specific measures of error
which can help users understand how accurate the forecast has been. Some
techniques seek to identify underlying patterns such as trend or seasonal
adjustments; others attempt to be self-correcting by including calculations about
past periods' error in future forecasts.
The code included here addresses several of the most common time-series
forecasting techniques, including naive/Bayes, simple moving average, weighted
moving average, exponential smoothing, and adaptive rate smoothing.
In the naive/Bayes approach, the current period's value is used as the forecast for
the upcoming period.
In a simple moving average, the prior n-number of values are averaged together
with equal weight to produce a value for the upcoming period.
Machine Learning Algorithm:
It proposes a system with two modules of forecasting: one for short-term
forecastings while the other one is being trained, and another one for long-term
forecastings. Machine learning borrows from the field of statistics, but gives new
approaches for modeling problems. The fundamental problem for machine learning
and time series is the same: to predict new outcomes based on previously known
results.
Modules:
Workload

Load the work, web and database server instances for the current observed
workload. Web infrastructures, the forecasting part does not represent a general
framework for workload forecasting.
Seasonal Study
The automatic building procedure for ETS does not choose seasonal models.
However, ARIMA can choose new seasonal and differenced models. After
modeling, likewise, if the original series are non-stationary, we also perform the
seasonal process on the differenced version of the series.
In seasonal time, load the product is to be added. Product is demand to the user.
Non Seasonal
In seasonal time, load the product is to be added. Product is more but the period of
time is less.
Time Series
It depicts the complete workload system series for the dataset hourly CPU load 2.
In the figure, we see an oscillating pattern with a period length of approx. one day,
which could indicate a seasonal pattern of daily fluctuation.
Products will be loaded a period of time. The time series will be calculated by the
system.
Forecasting:
Forecasting is the process of making predictions of the future based on past and
present data and most commonly by analysis of trends. A commonplace example
might be estimation of some variable of interest at some specified future date.

SYSTEM SPECIFICATION
Hardware Requirements:
System

: Pentium IV 2.4 GHz.

Hard Disk

: 40 GB.

Floppy Drive

: 1.44 Mb.

Monitor

: 14 Colour Monitor.

Mouse

: Optical Mouse.

Ram

: 512 Mb.

Software Requirements:
Operating system

: Windows 7 Ultimate.

Coding Language

: c#.NET

Front-End

: Visual Studio 2010 Professional.

Data Base

: SQL Server 2008.

Conclusion:
The Conclusion of Cloud Computing, as well as the relevance of having a
proper forecasting module, as part of a complete elastic system. We have seen
that workload forecasting in cloud platforms is a problem that needs a
preliminary stage to analyze and model properly the properties of the time
series. The effectiveness of this approach has been evaluated through its
application to workloads of four different cloud datacenters. The insights gained
from the empirical results have allowed to find the best model architecture, so

that the best forecasts from the viewpoint of the current state of the art in time
series forecasting have been achieved.

Você também pode gostar