Skip to content. | Skip to navigation

Informatik 5
Information Systems
Prof. Dr. M. Jarke
Sections
Personal tools
You are here: Home Projects Mobile Multimedia Cloud Computing

Mobile Multimedia Cloud Computing


Manager(s)
Research field Advanced Community Information Systems (ACIS)
Funding BIT-RS
Status running

Cloud computing envisions the notion of delivering software services and customizable hardware configurations to public access, similar how public utilities (electricity, water, etc.) are available to the common man. The cloud abstracts infrastructure complexities of servers, applications, data, and heterogeneous platforms, enabling users to plug-in at anytime from anywhere and utilizes storage and computing services as needed at the moment. The goal of our mobile multimedia cloud (i5Cloud) is to provide infrastructure as a service (IaaS) and platform as a service (PaaS) for diverse services and applications in the domain of (mobile) multimedia and large-scale social network analysis.

Realization

Our i5 multimedia cloud architecture consists of four layers. From bottom up, the infrastructure and platform layer focus on requirements from a technology perspective. The multimedia services layer considers the issues from a mobile multimedia perspective, while application layers demonstrate the user and community perspective.

IaaS

The infrastructure layer uses virtualization technology, namely Sun Solaris Containers, which decouples the software from the hardware. The virtual machines are grouped in realms which present boundaries between different computing resources. We have three realms, i.e. processing realm for parallel processing over many machines, then streaming realm responsible for scalable handling of streaming requests, and general realm for running other servers such as XMPP or Web server. Virtualization, dividing a physical server into multiple isolated virtual environments with its dedicated resources is the core part of i5Cloud. With the help of virtualization, applications that work on i5Cloud can dynamically scale up and down according to the demand of computing resources.


i5 Multimedia Cloud Computing Architecture

Our cloud infrastructure is based on our Sparc T5240 Sun Server using Solaris Containers for virtualization which enables us to run up to 128 virtual instances. The i5Cloud framework provides web services to start, stop, persist, destroy and monitor virtual instances (Solaris Containers). These virtual instances operate on dedicated CPUs,  virtual/physical memory and storage resources which are easily configurable.

 Virtual instance's states

PaaS

Deltacloud plays an important role in the architecture, because its API layer enables cross-cloud interoperability on infrastructure level with other cloud providers, e.g. Amazon EC2. The Deltacloud Core framework is provided to assist in creating intermediary drivers that interpret the Deltacloud REST API on the front while communicating with cloud providers using their own native APIs on the back.

To ease the development of large scale data processing and storage applications we use the distributed computing framework Hadoop, which enables a cloud service similar to a platform as a service (PaaS). Hadoop ,an open source project which was inspired by Google's MapReduce and Google File System (GFS) papers handles large datasets effectively by its parallel processing model. The core of Hadoop are MapReduce, Hadoop Distributed File System (HDFS) and HBase. MapReduce is one of the most popular programming paradigms for convenient large-scale computing on commodity hardware.

Hadoop                      Delta Cloud

i5Cloud also provides PaaS such as creating and running a pre-configured hadoop cluster. i5cloud users can create hadoop clusters seamlessly by specifying the number of instances in the cluster and hardware profile of the the instances. 

SaaS

Cloud Video Transcoder is scalable hybrid cloud application which uses i5Cloud and Amazon Web Services (AWS)

. Video transcoding is known as expensive CPU intensive operation, therefore it is a suitable application domain for scalable cloud applications. The main idea behind the Cloud Video Transcoder to start and use extra instances in case of demand then terminate these instances. In more concrete sense, users upload multiple videos to the system; if the number of videos that are in the transcoding queue exceeds the number of free instances in the cloud, new instances are started and the videos are transcoded in parallel. When the transcoding queue gets empty these extra instances are terminated. In the cloud, the cloud video transcoding service transcodes the video into streamable formats and stores the different versions of the video.

 

 

The application also provides real-time monitoring by displaying detailed status of transcoding operations and overall CPU usage and CPU usages at individual instances of AWS instances. 

Cloud video transcoding - 1

 Cloud video transcoding - 2

Cloud video transcoding - 3

 

Personal Clouds

According to Gartner's definition, personal cloud allows consumers to seamlessly store, sync, stream and share using multiple connected devices such as smartphones, media tablets, televisions and PCs over the internet. The emergence of personal clouds reflects the "4S experience", consumers' desire to store, synch, stream, and share their content on regardless of device or platform seamlessly. Service providers must deliver an invisible content synchronization experience. Personal cloud services can't be tied to one specific device or platform. An invisible experience means it works on anything.

Personal Clouds for Learning

Links

http://www.oracle.com/[…]/index.html
http://hadoop.apache.org/

Research staff

Former staff

Publications

Document Actions
Project Publications

Dejan Kovachev, Ralf Klamma, Matthias Jarke

CAELUS: Cloud Architecture for Enabling Mobile Multimedia Services

Proceedings of GI Informatiktage 2014, Lecture Notes in Informatics (LNI)

Dejan Kovachev, Ralf Klamma

Supporting Practices in Professional Communities Using Mobile Cloud Services

Kuan-Ching Li, Qing Li, Timothy K. Shih (eds.): Cloud Computing and Digital Media: Fundamentals, Techniques, and Applications, Chapman and Hall/CRC, 2014

Florian Stegmaier, Harald Kosch, Ralf Klamma, Mathias Lux, Ernesto Damiani

Multimedia on the Web - Editorial

Florian Stegmaier, Harald Kosch, Ralf Klamma, Mathias Lux, Ernesto Damiani: Multimedia on the web - editorial, Multimedia Tools Appl. 70(2): 821-826 (2014) (2012 Impact Factor: 1.014). DOI:10.1007/s11042-013-1729-9

More publications…