ViewVC Help
View File | Revision Log | Show Annotations | Revision Graph | Root Listing
root/i-scream/projects/cms/documentation/papers/wbem.txt
Revision: 1.1
Committed: Mon Oct 30 00:25:00 2000 UTC (24 years, 1 month ago) by ajm
Content type: text/plain
Branch: MAIN
Log Message:
A paper on the WBEM and CIM standards, added for the first time
Dunno if this is all you want to know.

File Contents

# Content
1 WBEM, CIM and the world of defined standards...
2 ===============================================
3
4 ajm4, 29/10/2000
5
6 What is all this?
7 -----------------
8
9 Web-Based Enterprise Management (WBEM) is an initiative
10 based on a set of management and Internet standard
11 technologies developed to unify the management of enterprise
12 computing environments. WBEM provides the ability for the
13 industry to deliver a well-integrated set of standard-based
14 management tools leveraging the emerging technologies such
15 as CIM and XML.
16
17 The Common Information Model is a common data model of an
18 implementation-neutral schema for describing overall
19 management information in a network/enterprise environment.
20 CIM is comprised of a Specification and a Schema. The
21 Specification defines the details for integration with other
22 management models (i.e. SNMP's MIBs or the DMTF's MIFs)
23 while the Schema provides the actual model descriptions.
24
25 The above are quite clearly lifted from the glossary on the
26 DMTF website. Mainly because they give the best definition.
27
28 How can we use them?
29 --------------------
30
31 One of the main topics of conversations that seems to have
32 come up time and again is how are the "clients" going to
33 communicate with the main system. CIM and WBEM provide a
34 well defined standard of passing objects of data throughout
35 a system. This means that the data undergoes no translation
36 at any stage, the "client" sources the data, wraps it in XML
37 in the CIM Schema and passes it to the main system. This
38 can then be passed so alter mechanisms, or can even be
39 attached a CIM object that has a metaschema that allows
40 alerts, though this is very advanced. The main point is
41 that it allows a universal representation of data that can
42 be interpreted by all parts of the system as well as
43 external systems.
44
45 Why should we use them?
46 -----------------------
47
48 Once of the communications methods frequently suggested by
49 the group is the use of SNMP. SNMP has been around many
50 years now and is well established. However it is a heavy
51 and dated protocol, heaving under extensions and
52 modifications that have been made to it over the years. The
53 DMTF or Distributed Management Task Force have been trying
54 to use SNMP as a basis for dragging systems management into
55 the new millenium. They have gathered together people who
56 understand the shortfalls of SNMP and have (and still are)
57 developed a system that will in the long term replcae SNMP.
58
59 Many companies have already taken and implemented the
60 standards such as WBEM that DMTF have produced. Indeed, few
61 motherboards today are not DMI (Distributed Management
62 Interface - another DMTF standard) compliant.
63
64 As the main part of our backend system is possibly (well,
65 probably) going to be predominantly Java based, it is
66 perhaps good to mention that Sun have been on the DMTF
67 bandwaggon for some time. This has two main advantage, 1)
68 There is a Solaris SDK for WBEM - quite what use that will
69 be I don't know 2) There are Java API extensions in the form
70 of JMX - java mangement extentions, which implement the WBEM
71 and CIM standards. Microsoft, Cisco, IBM and many others
72 have all implemented WBEM in various forms.
73
74 So where do we start?
75 ---------------------
76
77 I don't know.
78
79 I have read the introductory information, seen some
80 presentations and read a tutorial and some white papers. It
81 is big, but no bigger than a standard such as SNMP. The
82 system should certainly be investigated by all members of
83 the group, we should endevour to help each other understand
84 it. However it may be better to simply ignore this until
85 such a stage that we feel we are "on top of" the project and
86 that we can spare time to investigate this route.
87
88 One of the things that has been discussed is "protocols",
89 these are unimportant at this early stage. Indeed we could
90 implement the framework without really bothering about the
91 protocol (just use some simple method). As long as the data
92 encapsulation is seperated enough from the function then we
93 have no real problem. The system can operate, calls can be
94 fired and the system can work, then at a later stage the
95 classes that deal with the data communication and storage
96 can be replaced and updated (in line with the implementation
97 phases). However this still means that a concertated effort
98 needs to be made into deciding *what* data will be used,
99 *how* it will be used and what interfaces will the various
100 components need (ie, accessors, mutators and other methods -
101 think java interfaces, not GUI).
102
103
104 Online Resources
105 ----------------
106
107 There is little more that can be said on this subject until
108 a stage when we choose to pursue it further. Some useful
109 online resources are:
110
111 The home of all the information that is specific to the
112 standards group.
113
114 http://www.dmtf.org
115
116 Sun's Java API that implements the standards (JMX).
117
118 http://java.sun.com/products/JavaManagement/
119
120 Sun's central WBEM site
121
122 http://www.sun.com/software/solaris/wbem/
123
124
125 About
126 -----
127
128 This document was written by Alex Moore [ajm4@ukc.ac.uk] for
129 use by the team working on a 3rd year Computer Science
130 project called "i-scream". More details can be found on the
131 project website;
132
133 http://www.i-scream.org.uk
134
135
136
137