Adapting content for mobile devices in heterogeneous collaboration environments

Educación y Pedagogía

8 pages

Please download to get full document.

View again

of 8
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Abstract The portability of new miniaturized devices, together with their ability to connect conveniently to networks in different places, makes mobile computing possible. Recent advances that have led to increasing bandwidths, for wireless
   0   Adapting Content for Mobile Devicesin Heterogeneous Collaboration Environments Sangmi Lee 1, 2 , Sunghoon Ko 1 , Geoffrey Fox 1, 3  Community Grid Computing Labs, Indiana University 1  Department of Computer Science, Florida State University 2  Department of Computer Science, Indiana University 3, {suko, gcf} Abstract The portability of new miniaturized devices, together with their ability to connect conveniently to networks indifferent places, makes mobile computing  possible. Recent advances that have led to increasing bandwidths, for wireless communications, enable mobile users to utilize a vast range of sophisticated services such as real-timecollaborative services. Garnet Collaborative system is a universal accessible collaborative environment supportingsynchronous communication and universal accessibility from a wide range of devices. Garnet Message System Micro Edition (GMSME) is developed to integrate heterogeneous mobile devices into the Garnet collaborativesystem. Garnet integrates desktop in traditional LAN environments, wireless LAN based mobile devices, and pervasive 3G communication technology seamlessly. In this paper, we describe the process of adapting content inGMSME. GMSME provides application-aware transcoding, to tune the shared resources, for different collaborativemodels. Also, it optimizes shared resources for each heterogeneous client environments. Key words : Content Adaptation, Transcoding, Collaboration, Mobile Computing 1.   Introduction The availability of miniaturized devices and wireless communications has made mobile computing feasible.An ever more mobile workforce, and the computerization of inherently mobile activities is driving a need for applications to be integrated with traditional distributed systems. Mobile cellular telephones are widely availablethese days. A handheld computer is integrated with such a telephone is called as a smartphone. Currently wirelessnetwork service vendors have introduced a wide bandwidth telephone network, 3G communication [2], and itenhances the possibility of adapting a smartphone as a client in traditional distributed systems. Existing wirelessLAN technology is already equipped to  Personal Digital Assistants (PDAs) and laptops, and enables  pervasivecomputing  with its wide-bandwidth network communication [3]. Pervasive or ubiquitous computing is a newemerging computing style which adapts various computing devices throughout our living and working spaces. Thesedevices include PDAs, smartphone, traditional desktop, wearable computers and so on. These devices coordinatewith each other and network services seamlessly.We address the issue of pervasive computing for our collaborative system, Garnet  [4]. We have beeninvestigating design of an adaptive collaborative system supporting multiple heterogeneous devices. Moreover, wegeneralize it to universal collaboration – the capability of multiple users to link together over the web with disparateaccess modes. Despite innovative advances in computing capability and wireless communication services, mobilesystems are typically slow, unreliable, and have unpredictable temporal characteristics. Further, the user interface isclearly limited. The design of distributed mobile applications needs to identify the practicalities, reliability, and possibilities of continuous interaction and integrate synchronous and asynchronous collaboration.While adapting new devices, one should consider various factors: network communication mechanisms,computing capabilities, display capability, etc. We developed the Garnet Message System Micro Edition (GMSME)[20] to provide adaptability to the Garnet system. The major functionalities of GMSME are – support for heterogeneous wireless network communications, the manipulation and optimization of content, support for collaborative environment, and finally the integration of the 3 rd party applications for mobile devices. In this paper,we focus on how we tune the shared content for the limited capable mobile devices according to the characteristicsof specific collaborative features.Originally, transcoding technology is considered for converting multimedia from one format to another format preferred by specific devices [12]. We utilize various transcoding technologies to adapt shared content basedon the type of collaboration. Here transcoding technology includes image resizing, converting image formats,   1 compressing data, and transform technology such as the use of stylesheets. There have been a number of approachesto accommodate mobile devices by transcoding technologies [25-29]. However, different collaborative features mayneed different transcoding technologies according to their unique functionality. In this paper, we will explain our approach based on two different and popular resource sharing model, shared display and shared export.In shared display, one shares the bitmap display and the state is maintained between the clients bytransmitting the changes in the display. Meanwhile, the shared export model filters the output of each application toone of a set of common formats and builds a custom shared event viewer for these formats. This allows a singlecollaborative viewer to be leveraged among several different applications. Document formats such as W3C’sScalable Vector Graphics (SVG) [15] or Adobe, Inc.’s Portable Document Format (PDF) [10] are particularlyinteresting and support of collaborative viewers is a major advantage of Garnet System. With the scalability of vector graphics, and separation of user presentation from the master’s content, shared export provides moreflexibility for scientific visualization or geographical information systems. The constraints of “real-timesynchronous collaboration” for both methods, implies a delay of 10-100 millisecond for each participant [1]. Inshared display method, the time constraint is the most important factor to perform its illusion of collaboration.Basically shared export also requires this “real-time” constraint; however the graphical quality has higher prioritywithin the range of time delay for synchronous collaboration. We investigate technology for adapting content thatcan maximize the collaborative features and optimize network bandwidth.The rest of this paper is organized as follows: Section 2 reviews other related works in the area of pervasivecomputing and collaboration systems. In section 3, we describe the data flow in Garnet collaborative system. Wealso discuss utilizing transcoding technologies in GMSME to fit the shared content for limited display system insection 4. We also describe application-aware transcoding used in GMSME, with our resultant data. The conclusionand our future directions are discussed in section 5. 2.   Related Work  Adapting new devices and emerging technologies innovatively to collaborative services are described in [5-7]. Rendezvous [8], GroupKit [9] and several groupware toolkits utilize the model-view so that services cangenerate different views for different users. However, since the situation is greatly simplified by using a centralizedarchitecture, they were not able to support the diversity in devices accessing the service. One of the early efforts toadapt PDAs to work with conventional desktop computers is Carnegie Mellon University’s Pittsburg Pebbles PDAProject [6]. Pebbles is designed to communicate with PDAs through the communication server, PebblesPC, andevery message is conforms to the Pebbles protocol defined in their header files.Transcoding is one of the most popular ways to tune content from a service provider. Transcoding is thetransformation that converts the multimedia object from one form to another, frequently trading object fidelity for size [25], and is used to convert image or video formats (reduction resolution or compress data). A classicalapproach from ORL AT&T Lab’s VNC [13] allows user to customize encoding methods. As extended versions of VNC for mobile devices, there are Harakan Software’s PalmVNC [7], Nokia’s Java VNC viewer [14]. Theseapproaches implemented the client side of the VNC protocol for PDAs and wireless phone. This They enabledmobile device to be a thin client of a traditional desktop coupled with the sharing of master’s view and the executionof user events.Transcoding technology is also used to fit document and graphics files to the unique constraints of mobiledevices and other Web-enabled products. A number of distributed services use transcoding technology to generatedocuments for their heterogeneous clients [25-29]. The Apache Cocoon [32] project allows automatic generation of HTML, PDF, and WML files through the processing of statically or dynamically generated XML files using XSL[33] and XSLT [11]. The idea of transcoding is also adapted to Web service architectures. Duke University’s QualityAware Transcoding project [30] investigates differentiated Web services, which enables the Web services and Webservers to manage their available bandwidth with its quality aware transcoding. IBM introduces WebSphereEveryplace Access [12], which supports developers and administrators to utilize the transcoding technology byaccessing portlet which performs transcoding. Transcoding technology is developed as a component of WebSphere,which is IBM’s Web service infrastructure [24]. WebSphere Everyplace Access is designed to perform itstranscoding process in their individual portlet. This approach is different from existing distributed systemssupporting the transcoding process for heterogeneous devices, because it enables the transcoding process to beseparated from the proxy architecture. It also enables content providers to provide high-quality transcodingtechnologies on the server side.   2 3.   Dataflow in the Garnet Collaborative System Garnet system enables synchronous/asynchronous communications and provides a collaborationenvironment for heterogeneous clients. The Garnet system deploys a conventional shared event model for collaboration. Garnet ensures that objects are shared properly between collaborating clients with messaging to propagate changes to the object’s state. Hand-held devices, mobile phones, as well as conventional desktops can joinin one collaboration session of Garnet system. Every message delivered to mobile devices by GMSME. GMSME isa proxy-based service unit for mobile devices [20].As depicted in Figure 1, shared graphical content from Garnet is processed to fit within the display spacefor each mobile device based on the user and device profile. There are four major data processing stages to tune thegraphical content: using stylesheets, resizing the graphic, graphic format converting, data compression. Each processing stage can operate alone or can be cascaded to facilitate processing for a specific device. There is also animage renderer for generating image from graphical resources described in graphical markup languages, such asSVG document. Note that this dataflow is deficient in the sense that there must be a clear flow not only from the GarnetCollaborative Service to mobile user, but also back again to process the user interactive event. A collaborative eventin a specific device should interoperate with different types of devices correctly. This can be quite complicated andit is not clear how this is achieved in general with various transformations. For this reversibility problem, thesrcinal document is kept in transcoding unit itself, which allows us to access the document from each processingstage for generating the corresponding interactive event for different devices. 4.   Application -aware Transcoding in Garnet Here we discuss transcoding technology utilized in Garnet to support mobile devices with limitedcapability. Resource sharing in real-time is performed with events recording state changes transmitted from a“master” instantiation to replicas on other clients in the same session. The resources include visualizations, web pages and Microsoft Word documents and are shared with  shared export  or   shared display mechanisms on the PDA.Shared Display is the simplest method for sharing documents with the frame-buffer, corresponding to either a window or the entire desktop, replicated among the collaborating clients. Shared display does not allow significantflexibility; for instance, different clients cannot examine separate viewpoints of a scientific visualization. Moreflexible sharing is possible by sharing object state updates among clients with each client being able to choose howto realize the update on its version of the object, a process known as the shared export mechanism [1].The transcoding technology should consider the characteristics of the collaborative feature. Shared displayrequires fast and efficient propagation upon the master’s change of display. Thus reducing the size of data over thewireless network is critical. On the other hand, shared export provides more flexible visualization and needs toensure quality of the graphic and user interactivity to provide a user’s preferred view.The transcoding process in GMSME inherits Handheld Message Service’s (HHMS) [20] transparentapplication-awareness. When the master of the session changes his or her collaborative feature, the mobile users donot have to initiate new features or fix their profile to adapt new resource. Also the underlying transcodingtechnology is performed according to the changed features. In this section, we will describe the transcodingtechnologies used in our collaboration prototypes, shared display and shared export in detail.The Garnet system is developed with pure JAVA solution. The client application running on mobiledevices are also based on JAVA techniques, SUN’s J2ME platform [21]. GarnetCollaborativeService Transcoding    U  s  e  r  a  n   d   D  e  v   i  c  e   P  r  o   f   i   l  e StylesheetsResizingGraphic FormatTransformationCompression    I  m  a  g  e   R  e  n   d  e  r  e  r   Figure 1. Dataflow for adapting mobile devices   3 4.1 Transcoding in Shared Display Since Shared Display is sharing documents with the frame-buffer, modest client dependence is possiblewith mobile devices, for example receiving a reduced size image. Some collaboration systems support remotemanipulation with user interactions on one machine holding a replica frame-buffer transmitted to the instanceholding the srcinal object. This is an important capability in help desk or consulting applications, similar tosituations that occur frequently in the debugging of code. As this works for all applications without modifying them,this is the basic shared document mechanism in Garnet System [4].The Garnet system processes image captured from the master’s framebuffer in three ways. First, it resizesthe images based on the device profile. Table 1 shows a resultant data from a test session with Microsoft Power  point. The resolution is customizable on the client side. Second, Garnet system supports graphic formattransformation for specific environment, for example  Mobile Information Device Profile (MIDP) from Sun supportsonly raw bitmap data and the PNG (Portable Network Graphics) [34] graphic format. To deal with this case, Garnetsystem generates PNG image for MIDP equipped smartphone client. Finally, it compresses image data to savenetwork bandwidth. Garnet compresses image data with Huffman [35] and LZ77 [36] algorithm. For raw bitmapdata, comprising of 8 bit of RGB and alpha, Garnet eliminates alpha bits which represents transparency of image,and reduces the communicating data size to 25%. Table 1 presents the actual data size transformed via a wirelessLAN network or 3G communication service as well.HardwareSpecificationDesktop PC1280 x 1024LANSUN J2SE PDAs (iPaq 3900)240 x 320Wireless LAN 802.1bSUN PersonalJAVA Smartphones (Treo300)160 x 1603GWirelessCommunicationSUN MIDPImageImage Size (pixels) 1106 x 930 553 x 465 120 x 100Transmitted data size(KBytes)4017.9 KB 50.9KB 1.7 KBImage Format Bitmap image Bitmap image PNG image Table 1. The resultant data from test session of Garnet shared display4.2 Transcoding in Shared Export Since shared export allows the users to access more abstract stages of an object, it provides significantflexibility with view points based on user preferences [16]. For instance, different users can browse a display datathat is generated from identical data sources but rendered in different way. This is very time consuming to develop if one must do this separately for each shared application. The shared export model filters the output of eachapplication to one of a set of common formats and builds a custom shared event viewer for these formats. Thisallows a single collaborative viewer to be leveraged among several different applications.One of major advantage of Garnet System is its supporti for a collaborative viewer based on W3C’s SVGformat. Scalability implies that each client can resize and scroll while preserving correct positions of pointers andannotations for their various resolutions. As depicted in figure 2, the benefit of the vector graphics format is itsscaling/resizing ability. For mobile devices which have limited display size, this ensures flexible resource access based on user preferences. SVG is useful as it is already available for Adobe Illustrator [14] and both PowerPointand Macromedia Flash are exportable to this syntax. Currently there is a Flash (which is a binary 2D vector graphicsformat) to SVG converter [37] from the University of Nottingham while’s OpenOffice [38] exportsPowerPoint to SVG.   4  Another advantage of SVG is its formatting effects with stylesheets. SVG inherits XML’s stylingtechnology such as the use of CSS syntax and properties or XSL. Figure 3 shows the different output imagesrendered with different stylesheets. The advantage of styling with stylesheet is flexibility in reformatting images. Inaddition to color adjustment, figure3(b) is transformed by CSS stylesheet, which redefines the width of a brushstroke to be wider and the text style to more recognizable.To provide individual presentation to each user, Garnet allows every user to own their object instance inshared export. Figure 4 is the object flow of the SVG shared export for each user. Image rendering andtransformation are performed in GMSME to reduce the workload in mobile devices. Eventually the ready-to-useimage is delivered to PDAs or smartphones. There are several graphic formats supported in mobile devices. SUN’sMIDP supports only PNG and raw bitmap graphics, while SUN’s Personal Java runtime environment supports JPEGand raw bitmap graphics. Thus, we provide format conversion for specific mobile environments. Some of thegraphic formats include data compression mechanisms; however, raw bitmap image data needs additional datacompression to better utilize wireless network bandwidth. Data compression capabilities can be fine tuned byindividual users based on their needs.Since shared export is designed for more flexible and high quality resource sharing, stringent timingconstraints are not the overriding factors. With the scalability of vector graphics, shared export provides maximizedresource sharing to users. The users are able to browse the best quality of image supported by their devices. 5.   Future work and Conclusion For future work, the process of adapting content will be developed as a component of a Web service.Performing transcoding in proxy-based middleware may have disadvantages in a large user group. As we presentedin Table 1, the actual data delivered to mobile devices are much smaller than the srcinal data from master client.However, the transcoding functions should download whole amount of data to process it, which can incur significant   (a) Bitmap Image (b)Vector Graphics (a) (b) Figure 2. 400 % Zoom In Images   Figure 3. Styling with CSS for Black and White PDAs   SVG DocumentRenderingImage FormattingData CompressionStylesheetUser ProfileUser Event   Fiure 4. Obect flow in SVG Shared Exort in Garnet
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks