Dr. Vrbsky's research is in the area of data intensive computing, cloud computing, real-time database systems, mobile databases and database security.
Cloud and Cluster Computing (with our private cloud Fluffy and our green cluster Sage)
Data Management in Cloud
Computing:
with Md. Ashfakul Islam
Data sizes are increasing exponentially everyday, which makes data management
applications potential candidates for deployment in a cloud.
The
main features of cloud computing: scalability, availability and reliability are
achieved by sacrificing consistency.
While different forms of consistent states have been introduced, they do not
address the needs of many database applications.
A large part of the data management market is transactional data management, and
the maintenance of ACID properties is the primary obstacle to the implementation
of transactional cloud databases. As a result, many key questions need to
be answered regarding consistency maintenance before transactional cloud
databases can be a viable solution for the user.
The goal of our work is to provide low cost solutions to ensure data consistency
in the cloud.
Load Balancing in Cloud
Computing:
with Michael Galloway
As local clouds become more common in organizations, they will realize the
benefits of scaling, even if only temporary, to public cloud sources.
Resource allocation in a private cloud architecture is a challenging task.
There
have been many load balancing algorithms proposed by others, generally
considering CPU, memory, network, or a combination of these resources.
These propositions do not work as well when the workload consists of a large
number of I/O intensive requests.
Simple, static load balancing algorithms, such as random distribution or
modulus-based round robin are
not the most efficient for use with heterogeneous cloud architectures
Proposal
of new, dynamic, resource-balancing algorithms that are platform independent and
are efficient based on the client’s are needed.
Our private cloud Fluffy will be used to
test our proposed new algorithms.
Security in Cloud Computing:
with Kazi Zunnurhain
Cloud computing has been envisioned as the next generation
architecture of IT Enterprises. Cloud computing
moves the application software and databases to large data centers, where the
management of the data and services may not be fully trustworthy. This attribute
poses many new security challenges which have not been well understood yet.
We are investigating the possible security
attacks on clouds including: Wrapping attacks, Malware-Injection attacks and
Flooding attacks, and the accountability needed due to these attacks. The focus
of this work is to identify and describe these prime attacks with the goal of
providing theoretical solutions for individual problems and to integrate these
solutions. The result will be to provide
additional security in the context of cloud computing.
Green Grids:
with Karl Smith and Ming
Lei
This project involves three facets: 1) Designing energy efficient strategies for data
intensive computing systems that minimize disk storage and data transmission,
and maximize the length of CPU bursts. Our
previous results for data replication and CPU scheduling indicate that
for data intensive computations, opportunities exist for energy savings beyond
what is seen in numerically intensive computations.
2) Implementing these energy efficient strategies on a prototype green cluster
SAGE at
the University of Alabama.
We will use each of these clusters to simulate Grid behavior.
Results will indicate if the proposed strategies provide a benefits in
terms of running time and power consumed.
Database Security
Virtual Passwords:
With Dr. Ming Lei and Dr. Yang Xiao,
While people enjoy the convenience of shopping online, ATMs and many other
online services, pervasive computing brings many risks. We propose a
virtual password concept that involves a small amount of computing (human or
machine) to secure user's passwords. We adopt such strategies as
randomized linear generation functions, codebooks and secret little functions to
secure users' passwords. We analyze the proposed schemes and have found
them to defend against the three attacks of: phishing, key logger and
shoulder-surfing attacks.
Belief-Consistent Multilevel Secure Data Model: With
Dr. Nenad Jukic
Multilevel relations, based on the current multilevel secure (MLS) relational
data models, can present a user with information that is difficult to interpret
and may display an inconsistent outlook about the views of other users.
Such ambiguity is due to the lack of a comprehensive method for asserting
and interpreting beliefs about information at lower security levels.
We identify different beliefs that can be held by users at higher security
levels about information at lower security levels, and introduce a mechanism for
asserting beliefs about all accessible data.
This mechanism provides every user with an unambiguous interpretation of
all viewable information and presents a consistent account of the views at all
levels visible to the user.
This model is also applicable for access and flow control in virtual
organizations.
Mobile Databases
Mobile Real-Time Databases:
With Dr. Yang Xiao and Dr. Ming Lei
This project addresses the issue of satisfying real-time constraints of databases in a mobile environment. Real-time database systems can be found in such applications
as air traffic control, industrial control, and programmed stock trading. Many
of these applications would benefit from a mobile environment that allows
wireless access to data, but the characteristics of a mobile environment, such
as disconnection, add to the challenge of satisfying the constraints of a
real-time database. We are designing a collection of broadcast strategies for the mobile
environment which contain both requests with deadlines and requests without
deadlines.
We do not consider these issues in isolation, but
consider our
strategies within the limited resources available in a wireless network
environment.