Scalable Cloud for Urgent Massive Email Delivery
Mauricio Alejandro Gómez Morales, Pham Phuoc
Hung, Md. Motaharul Islam, Eui-Nam Huh
Department of Computer Engineering
Kyung Hee University
Seoul, South Korea
{mgomez, hungpham, motahar, johnhuh}@khu.ac.kr
Seung-Bae Yang
BiziSolution
Seoul, South Korea
tdory@bizisolution.co.kr
Abstract— When there is the necessity to send an email
message to a big list of email destinations and following the valid
standard Simple Mail Transfer Protocol (SMTP), it takes too
much time to be delivered, because this message is sent one by
one to each of the email destinations. If we talk also of an
important message that all the recipients are in need to receive it
in a timeless way, converting this need as a Real Time System,
that will let them being able to take the corresponding immediate
actions and specifically we can say that all the destinations of this
email are really looking forward of this message, with this we can
say that it will never be taken as a SPAM email. With all the
mentioned above, in this paper we propose a solution by
provisioning resources, and escalating the exact infrastructure
needed for the time that is going to be used to deliver the message
within a small time requirement defined by a Service Level
Agreement (SLA). In this manuscript, we designed the
procedure to determine the structure, not only the VMs needed
to achieve the requested time, and taking into account the
contention problem and we also made a cost/efficiency
comparison with the different parameters available for the
customer.
Keywords— Real time systems, cloud computing, optimal
resource usage, SLA, SaaS
I. INTRODUCTION
As time passes, for human kind the necessity to obtain
information on time becomes more important. In terms of
important messages, we do not only consider the business
sector (advertisement that is considered by some people as
SPAM) but also the government, global organizations and
other matters of communication. The most common
communication way over internet is email, not only because of
its simplicity but also because of its effectiveness to spread
messages, news, advertisements and others around the world
with almost no effort and investment. When the message is
needed to be sent to a huge amount of destinations, it is
classified as Massive Email.
Cloud computing is a model for enabling ubiquitous
convenient, on-demand network access to a shared pool of
configurable computing resources (e.g., networks, servers,
storage, applications, and services) that can be rapidly
provisioned and released with minimal management effort or
service provider interaction [1]. The most important idea
behind Cloud Computing is scalability, and the key technology
that makes that possible is virtualization [2], [3]. Recently
cloud computing is being used to scale the services that need
resource on demand, and to de-allocate those resources when
are no needed anymore, this, improves the efficiency of
resource usage. Precisely, one of the fields that can be
enhanced is email sending and more specifically speaking the
massive email communication.
Real Time Systems are computer systems that must react
within precise time constraints to events in the environment
[4]. With this, when the message to various destinations are
needed to be delivered before a requirement of a Service Level
Agreement (SLA) the system will be considered as Real Time
System.
Adoption of cloud resources for this purpose is not
investigated in a big manner until these days. A bottleneck
could be noticeable when there is communication resource
sharing. This bottleneck resource is measured in terms of
available bandwidth of a network path P that for [5] is the
maximum throughput that P can provide to a flow, without
reducing the throughput of the cross traffic in P. In Figure 1 is
shown the problem that we are talking about.
This is how our main contribution is a model that
dynamically calculates the combination of a network
/datacenters-VMs needed to accomplish a specific deadline of
delivering a massive email, taking into account and solving net
contention problems.
The rest of this paper is organized as follows: Section two
surveys the related works on cloud computing and email
delivery system, the third section presents the problem
formalization, on fourth section simulation results are shown,
and in the fifth and last section we present the conclusion and
future work.
II. RELATED WORK
In [6] is designed and implemented an email-based personal
cloud storage that aggregates back-end services by establishing
a RAID-like system. Also in [7] is proposed architecture to
eliminate duplication data in the storage of email systems to
save storage space and improves performance of email server.
On the other hand, the majority of the cases in research area of
the mixture of terms of email and cloud, are more focused in
security and privacy issues, of any kind, like [8] where they
propose a scalable cloud spam analysis framework to address
the shortcoming of deploying a server based spam filtering
toolkit into a commercial cloud computing environment. Also
978-1-4799-0604-8/13/$31.00 ©2013 IEEE