Information Technology Reference
In-Depth Information
Themiddlelayeroftheproposedmodelisan
infrastructure requirement on-
tology
. This layer describes provider-agnostic infrastructure constraints that are
needed to deliver the application requirements. In going from the top to the
middle layer, high-level
domain-specific
requirements are mapped to
infrastruc-
ture level
requirements. The bottom layer in the model is the
resource
layer
which specifies the resource capabilities offered by various cloud providers. This
paper focuses on the
domain-specific
and
infrastructure requirement
layer. The
resource
layer has been widely investigated elsewhere [2,8].
3.1 Domain-Specific Ontology
A domain-specific ontology can be used to capture high-level application con-
straints. The ontological layer is application-centric, focused on user needs and
expressed using domain specific terminology. Two examples of domain-specific
ontologies are given in order to illustrate the model.
Media transcoding
is the process of converting media files (video or au-
dio) from one format to another. Transcoding is computational intensive and
requires high storage and fast bandwidth [6,10]. Often users impose a budget
for the provisioning of transcoding infrastructure. Consider a media company
that broadcasts a series of animation videos. The video sources use
avi
format
and are made available 5 hours before the broadcast schedule. For certain ap-
plications, these need to be transcoded into
windows media
format at a frame
rate of
30 frame per second
and delivered to
Windows Phone
devices via
http
streaming
. The company has a budget of
£100
for the transcoding operations.
The application's requirements may be specified in a high-level notation as:
Video conversion : AVI to Windows Media
Mobile encoding : Windows Phone
Delivery deadline : 9am next morning
Encoding features : frame-rate conversion; http adaptive streaming
Budget :
£
100
More generally, the following domain-specific ontology is used for specifying me-
dia transcoding requirements:
Budget requirements
specify monetary constraints for running a transcoding
task. These can be specified as the maximum amount that a user is prepared
to spend per day or per hour.
Format requirements
specify the container format of the source and
transcoded media; for example, transcoding a video from
avi
format to
flv
format.
Codec requirements
are the audio or video codecs of the media; for example,
mpeg4
,
h264
,
mp3
,
aac
.
Device requirements
refer to the destination devices that the transcoded
media will be played on; for example,
iPhone
,
Windows Phone
,
PC
.
Processing Filter requirements
are advanced video and audio filters, in-
cluding both pre-processing and post-processing filters; for example, frame
rate conversion, de-interlacing, watermarking, audio resampling, etc.