Information Technology Reference
In-Depth Information
Performance requirement
is constrained by performance-related restric-
tion.
Latency performance
is constrained by
latency restriction
(expressed
via
hasLatency
property).
Bandwidth performance
is constrained by
band-
width restriction
.
Bandwidth restriction
indicates minimum amount of
bandwidth required using the
hasMinBandwidth
property.
Resource requirement
is constrained by resource-related restriction.
Host-
ing environment
is constrained by
operating system restriction
which indi-
cates the operating system types (via
hasOS
property).
Hardware capability
is constrained by
hardware restrictions
,suchas
cpu core restriction
(via
hasMinCPUCore
property),
cpu speed restriction
(via
hasMinCPUS-
peed
property),
cpu architecture restrictioin
(via
hasCPUArchitecture
property),
RAM restriction
(via
hasMinMemory
property), and
storage
space restriction
(via
hasMinStorageSpace
property).
Software stack
is
constrained by
software restriction
which indicates the list of softwares or
services that need to be installed on the resource (via
hasSoftware
property).
Geographical requirement
is constrained by
location restriction
.
Lo-
cation restriction
is associated with
hasLocation
property that indicates the
location (country or data center location).
Compliancecoderequirement
is constrained by
compliance restric-
tion
,whichcanbe
standard code restriction
- contains the standard's code
format or
regulatory restriction
- contains the name of regulation.
3.3 Resource Ontology
The resource ontology, the bottom layer of the proposed model, defines the
properties of the resources offered by cloud providers. This layer has been widely
investigated elsewhere - see [2] and [8]. These existing ontologies can be applied
as the resource ontology in the proposed model. The mapping of infrastructure
requirements to the resource ontology can be achieved by using query language
[5] - this topic is outside the scope of this paper.
4 Translation from Domain-Specific Ontology
Here consideration is given as to how a domain-specific ontology can be trans-
lated to the infrastructure requirements ontology.
For the media transcoding example, application developers need to provision
an infrastructure which fulfils the transcoding requirements as well as balances
the budget and delivery time constraints. Transcoding requirements, such as
video format conversion, frame rate conversion or http streaming, indicate the
features or capabilities of a transcoding software that need to provide. Particular
software, such as FFmpeg or Rhozet, can be used to perform the transcoding
task. However, each software has different system requirements. For example,
Rhozet must be run on Windows operating system, whereas FFmpeg can be
run on Linux. Using the proposed multi-layer ontology model, domain-specific
transcoding requirements can be translated into software requirements, where