Professional Documents
Culture Documents
Authors
Name: Ekanath M. Patil & Aparna D. Kulkarni
th
7 sem- CSE
Hirasugar Institute of Technology, Nidasoshi
Email Id: ekanathl@gmail.com aparna2128@gmail.com
Mob: 9008616346 9008124392
I. INTRODUCTION
A. About cloud computing
Abstract - The term "cloud computing" is being hyped about a lot
these days, mainly in the context of the "future of the web." Cloud
computing refers to a style of computing where online resources and
applications are accessed by users through a Web browser, but the
actual software and data are stored on remote servers (or “big
computers”). Users don’t need to purchase, manage or maintain the
technology "in the cloud" that supports them. In this paper, we focus
on mobile cloud computing, the newest application of cloud
computing, which is recently seeing explosive adoption, since the
capabilities of mobile devices are enhanced by moving the computing
power and data storage away from mobile phones and into the cloud.
“Basically, it refers to an infrastructure where both the data TABLE 2: Specification of a commodity laptop and a desktop.
storage and the data processing happen outside of the mobile
device.”Some good examples of mobile cloud computing Computer CPU RAM
applications including mobile Gmail, Google Maps, and some MacBook Pro Laptop 2.5GHz 2-core 4GB
navigation applications. However, the majority of applications
Dell Precision T7400 3.3GHz 4-core 8GB
Their connectivities include 1Gbps Ethernet and Wi-Fi, and
they are frequently powered from the electric grid.
a) Primary functionality
outsourcing:
Computation hungry applications such as speech processing,
video indexing, and super-resolution are automatically split,
so that the user-interface and other low-octane processing is
retained at the smartphone, while the high-power, expensive
computation is off-loaded to the infrastructure, synchronously.
b) Background augmentation:
Unlike primary functionality outsourcing, this category deals
with functionality that does not need to interact with users in a
short time scale. This functionality typically happens in the Conceptually, this system provides a way to boost a
background, such as scanning the file system for viruses is off- smartphone application by utilizing heterogeneous computing
loaded and acts as virtual client. Even when the smartphone is platforms through cloning and computation transformation.
turned off, the virtual client continues to run background tasks. For doing so, our system (semi)-automatically transforms a
Later when the smartphone returns online, it synchronizes its single-machine execution (e.g., smartphone computation) into
state with the infrastructure. a distributed execution (e.g., smartphone plus cloud
c) Mainline augmentation: computation) in which the resource-intensive part of the
This category sits between primary functionality outsourcing execution is run in powerful clones. Figure 3 illustrates the
and background augmentation. Here the user may choose to high-level system model of our approach.
run a particular application in a wrapped fashion, altering the
method of its execution but not its semantics. Example is
Figure 3: The system transforms a single-machine execution
private-data leak detection. Unlike background augmentation,
(smartphone computation) into a distributed execution
mainline augmentation is interspersed in the execution of the
(smartphone and cloud computation)(semi)automatically.
application.
Augmented execution is performed in four steps:
d) Hardware augmentation: 1) Initially, a clone of the smartphone is created within the
This category is interesting because it compensates for cloud (laptop, desktop, or server nodes);
fundamental weaknesses of the smartphone platform, such as 2) The state of the primary (phone) and the clone is
memory capabilities or other constraints, and hardware periodically or on-demand synchronized; 3) Application
peculiarities. Just using faster hardware significant savings can augmentations (whole applications or augmented pieces of
be observed even while thrashing. applications) are executed in the clone, automatically or upon
e) Augmentation through request; and 4) Results from clone execution are re-integrated
multiplicity: back into the smartphone state.
The last category we consider is unique as it uses multiple Figure 4 Shows clone execution architecture for smartphones.
copies of the system image executed in different ways. This We achieve this by combining whole-system replication
can help running data parallel applications (e.g., doing through incremental checkpointing, (semi)-automatic
indexing for disjoint sets of images). This can also help the partitioning and invocation of augmented execution, and
application to “see the future,” by exhaustively exploring all coordination of computation between the primary (phone) and
possible next steps within some small horizon—as would be the clone. The system components are running inside the
done for model checking—or to evaluate in maximum detail operating system (OS). The Replicator is in charge of
all possible choices for a decision before making that decision. synchronizing the changes
IX. REFERENCES:
A. Websites: