Vous êtes sur la page 1sur 17

203

Conformance Testing for OSI Protocols *


Richard J. L I N N , Jr. 1. I n t r o d u c t i o n

National Institute of Standards and Technology, National Com-


puter Systems Laboratory, Gaithersburg, MD 20899 USA
Given that we have international standards for
communication protocols which define their be-
Abstract. In the early 1980's, research and development was
havior and encoding of messages, testing imple-
initiated on methods to test the developing International mentations of protocols should be done in a uni-
Standards Organization's Open Systems Interconnection form way and be relatively easy. This is the shared
(ISO/OSI) communications protocols. In 1983, ISO initiated goal of the standards bodies who define the proto-
work to develop standardized test methods. A tutorial overview cols and test methods, test laboratories, and
of the test methods and test notation named q'TCN which were
developed within ISO are presented. Issues regarding multi-
vendors who implement the standards in products.
layer test methods are still not completely resolved. These (Consumers of communication products just want
issues are identified and alternatives to ISO's methods are them to work together!) But given the complexity
explored. Application of the formal description techniques of Open Systems Interconnection (OSI) protocols
named ASN.1 and Estelle to multi-layer test systems is il- and the state of conformance evaluation methods,
lustrated. Concluding remarks summarize the status of current
practice in conformance testing and status of evolving OSI
"Is the goal a reality or a fantasy ... ?" The reality
testing standards. is that conformance testing is a difficult technical
issue fraught with problems. So, we start with a
Keywords. ASN.1, communications protocols, conformance discussion of the issues and problems which in-
testing, ferry methods, Open Systems Interconnection (OSI), eludes at least the following:
Estelle, test architectures, test methodology, Tree and Tabular
Combined Notation (TTCN).
- Initially, standards may be based upon a paper
design, and as a result, may contain errors,
omissions and ambiguities.
- Several options or alternative behaviors may be
allowed.
- Often, conformance criteria are not precisely
specified in a standard.
Richard (Jerry) Lima is a computer
scientist and manager of the Auto- - In a layered protocol architecture, is a stack of
mated Protocol Methods Program protocols to be tested as individual layers, as a
within the National Computer Sys-
tems Laboratory, at the National In- collective unit, or both?
stitute of Standards and Technology -Individual layer testing implies exposed in-
(NIST) in Maryland. Before joining
NIST, he was a senior systems en- terfaces must exist at each layer (interfaces
ginecr and managed groups support- might not be exposed in a product).
hag data communications, field en-
gineering, and systems engineering of -Multi-layer testing (a group of two or more
data acquisition and process control layers) raises issues of how layers should be
systems used in research laboratories
at Virginia Tech. grouped for testing; i.e., at which layers will
His recent research and development activities at the NIST interfaces exist and what services are actually
have focused on the formal description techniques for OSI
protocols named Estelle and Abstract Syntax Notation One implemented?
(ASN.1) and conformance evaluation methodology for OSI - A a result of decisions made regarding
s
protocols. He has contributed to international standards in
both areas. Linn received his B.S. degree in Forestry in 1972 grouping, multi-layer testing may restrict the
and an M.S. degree in Computer Science in 1981 from Virginia amount of testing possible on lower layers
Polytechnic Institute and State University.
within the group.
- Is the testing to be conducted:
* This work is a contribution of the National Institute for -locally (the test system has direct access to
Standards and Technology (formerly, the National Bureau
the interfaces of the implementation), or
of Standards) and is not subject to copyright.
-remotely (the test system has access to the
North-Holland implementation via underlying communica-
Computer Networks and ISDN Systems 18 (1989/90) 203-219 tions services)?

0169-7552/90/$3.50 1990, Elsevier Science Publishers B.V. (North-Holland)


204 R.J. Linn, Jr. / Conformance testingfor OSI protocols

- What test coordination procedures are to be of protocol data units (PDUs). Examples of the
used (if any) and what degree of coordination is dynamic aspects of testing are assessing the valid-
desirable a n d / o r acceptable? ity of messages sent in response to messages re-
- Are test coordination procedures to be auto- ceived and assessing error recovery mechanism.
mated? If so, what test coordination protocol The process of testing is outlined after ISO's test
definitions are to be used in an automated methodology is introduced.
environment? ISO has provided some answers in an evolving
- May requirements beyond those specified in a standard composed of five parts [11-13]. The scope
protocol standard be placed upon a supplier to of the standard includes definition of concepts
make a product testable? and terms, test methods, a test notation, and pro-
- How are tests to be specified: formas for use in protocol standards and test
- test cases, interpreted/executed by a soft- reports. It also defines requirements for suppliers
ware tool; of test systems and executable test suites, test
- by a suite of programs, each testing specific laboratories, and their clients. Since this work is
functions; intended to be applied by other working groups
- by a reference entity enhanced with features within ISO and CCITT, other answers are found
for testing? in the protocol standards themselves: annexes to
- Are the tests themselves to be standardized? If standards identify specific testing requirements
so, are test cases generated by manual, semi-au- and solicit specific answers from suppliers of OSI
tomated or fully automated methods? products. Additional answers are found in test
- How is test suite coverage to be measured, and suites being developed by ISO and CCITT for
how many tests are enough? individual protocols or groups of protocols (e.g.,
- How are test cases to be selected from a test what test methods will be standardized for a par-
suite given optional features in a protocols and ticular protocol).
valid implementation choices? This tutorial focuses on a subset of the techni-
- What information must be recorded during test- cal issues. It begins with an overview of early
ing to support verdicts rendered either manu- research and development. Then an overview of
ally or automatically? the ISO work on conformance evaluation method-
- What is the form and structure of the informa- ology presents a number of alternatives and
tion recorded? answers to some of the questions raised. Finally,
- What is the format of test reports: problems unanswered by current ISO work are
- what information must be reported; identified, and alternative solutions to the prob-
- what information must be treated as con- lems are proposed.
fidential;
- who owns the information; and
- when must a test report be made public? 2 . B a c k g r o u n d

- Are executable test systems to be standardized?


- Are testing procedures to be standardized? Within Europe in the early 1980's, formal col-
- How are mutual recognition of test results and laborative research efforts were initiated to estab-
test reports to be achieved? lish methods for testing implementations of com-
The issues range from technical to political and munications protocol, for conformance, as inter-
legal areas. The issues and the assumptions un- national standards were emerging under the
derlying the issues influence answers to the ques- umbrella of Open System Interconnection (OSI).
tions. Different resolutions lead to different ways Initial collaboration was between Agence de r l n -
to conduct conformance testing. Nonetheless, all formatique (ADI), Paris, France; Gesellschaft fiir
test methods attempt to assess the static and dy- Mathematik und Datenverarbeitung (GMD),
namic aspects of a protocol implementation with Darmstadt, F R G ; and the National Physical
respect to criteria specified in a standard. Exam- Laboratory (NPL), Teddington Middlesex, En-
pies of the static aspects of testing are assessing gland. Each research laboratory had industrial
the validity implementation choices regarding op- partners and focused upon a different aspect of
tions and services offered, and assessing encodings testing. ADI designed and implemented an X.25
R.J. Linn, Jr. / Conformance testingfor OSl protocols 205

tester; GMD developed a language oriented ana- some of the test methods described in the next
lytic tool for passive monitoring and error detec- section. However, this does not imply that all
tion for the ISO Session protocol; NPL developed aspects of OSI protocol testing are mature. Test-
a test system for a Network Service. Since the ing of lower layer protocols is the most mature
National Institute of Standards and Technology because they are older (e.g., X.25 and Transport);
(NIST) in the USA was developing a test system upper layer protocols are newer and present unique
for the ISO Transport Class 4 protocol, NIST was problems for two reasons:
invited to participate. Each of the tools developed (1) several layers of protocol may be embedded
were based upon different design architectures in a single product; and
and philosophies. By 1984, other researchers from (2) use of Abstract Syntax Notation One
Europe and Canada were invited to join the effort. (ASN.1) [7-8] to describe message syntax permits
Much of the early work is reported in the proceed- many encodings of the same message.
ings of the first through fifth IFIP workshops on The next two sections 1 introduce ISO's test
Protocol Specification, Testing and Verification methods and test notation (Tree and Tabular
[1-5]. Early work considered a variety of test Combined Notation--TTCN). The terminology of
architectures, test languages, and demonstrated ISO is adopted and is introduced in context.
the viability of several test methods, automated
generation of test sequences from formal specifi-
cations, and specification languages for protocols. 3. ISO's Single Layer Conformance Methods
In 1986, Bochmann [25] and his students surveyed
this work and identified tools for development of A brief overview of the methodology and
formal specifications, code generation from formal framework developed by ISO is presented. The
specifications and tools employed for testing. concepts and vocabulary used within ISO also
Results and terminology coming out of the provide the framework to discuss other work. The
initial research efforts were subsumed by ISO's ISO documents [11-13] and Rayner [40-42] pro-
work on conformance testing methodology. Ini- vide more detail. It is important to remember, this
tiated in 1983, it subsequently became a joint discussion focuses on models or a framework
effort with the International Telegraph and Tele- which are independent of implementation.
phone Consultative Committee (CCITT). A se- ISO's work defines a framework for the de-
quence of demonstrations between 1984 and 1988 scription of tests whose purpose is to provide
showed the progress of this work. Among the sufficient information to determine if an imple-
earhest demonstrations of OSI protocol was one mentation of a protocol conforms to a standard. A
at the National Computer Conference in 1984. judgment is made based upon static and dynamic
For six months prior to the show, the National aspects of an implementation. Static conformance
Bureau of Standards and General Motors Corpo- criteria focus on proper formation of data struc-
ration conducted testing on thirteen vendors' pro- tures, coding of protocol data units and condi-
totype implementations of the Transport Class 4 tional criteria which depend upon legitimate
and IEEE 802.3/4 protocols [34]. In 1985, the choices by an implementor about services offered
Industrial Technology Institute conducted tests on by the Implementation Under Test (IUT). Dy-
prototype implementations of the File Transfer namic conformance criteria focus on the behavior
and Management, Session, Transport, Connec- of an IUT as it interacts with another implementa-
tionless Network and IEEE 802.3/4 protocols in tion (or a test system).
preparation for a demonstration at the AUTO-
FACT'85 trade show [34]. In 1988, the Enterprise 3.1. The Local Method
Networking Event (ENE '88) demonstrated Mes-
sage Handling Systems (electronic mail) as well as The local method of conformance testing (Fig.
the other protocols operating over local networks 1) is important as the basis for a large body of
and global interconnections via wide area net-
works. Corporation for Open Systems demon- 1 D u e to overlapping subject matter, parts of Sections 3, 4 and
strated testing of OSI products at ENE '88. Pre- 6 are nearly identical to materials contained in another paper
ENE '88 testing employed tools which implement [38].
206 R.J. Linn, Jr. / Conformance testingfor OSl protocols

ceive a set of test events complementary to the set


Upper
Tester
of (Nt)-ASPs generated and received by the I U T
at its top interface. Similar assumptions are made
for the b o t t o m interface of the I U T and the lower
tester. The model in Fig. 1 comprises a test har-
ness around the I U T which coordinates the ac-
tions of the upper and lower tester. The roles of
the upper and lower testers are to stimulate the
I U T by exchanging test events at the top and
b o t t o m interfaces of the IUT. The lower tester is
also assumed to record (log) test events so that the
behavior of the I U T m a y be assessed and there is
evidence supporting p a s s / f a i l (or possibly incon-
clusive) verdicts.
Fig. 1. Local method.
Test cases provide the means of specifying the
actions of the upper and lower testers. Test cases
terminology and concepts that are subsequently specified in T T C N m a y be interpreted by a test
applied to other models. A basic assumption of system (manually translated, or compiled into
the local method is that exposed interfaces exist some executable form). The statements of a test
above and below the IUT. These interfaces serve case may make reference to named PCOs, (Nt)-
as points of control and observation (PCOs); i.e., ASPs, ( N b - 1)-ASPs, ( N ) - P D U s and their com-
points at which a real test system can control ponent fields. Thus, test cases define the actions of
inputs to and observe outputs from an IUT. the upper and lower testers which control and
Using ISO's conventions, the layer under test is observe test events exchanged at the upper and
referenced as the N-layer and the next lower layer lower interfaces of the IUT. In the local method,
as the ( N - 1)-layer. Protocol service definitions test-coordination procedures are expressed ab-
define abstract service primitives (ASPs) ex- stractly by the ASPs used to specify the actions of
changed at the top interface of a protocol entity. the upper and lower testers.
Abstractly, each protocol entity exchanges Note, even though the lower tester is below the
( N - 1)-ASPs with an underlying service provider. IUT, it functions as a peer protocol entity. The
ASPs comprise a logical set of test events (with lower tester exchanges N - P D U s with the I U T
parameters) which can be controlled and observed through its b o t t o m interface. Therefore, the lower
at the interfaces of an IUT. The set of test events tester may emulate normal behavior of an N-layer
exchanged at the top interface are denoted as protocol entity during testing, or it may inject
(Nt)-ASPs; t for top. The same is true of the errors to test for error recovery.
b o t t o m interface, and the test events are called In summary, the local method of testing as-
( N b - 1)-ASPs; b for bottom. Each protocol de- sumes a test harness around the I U T and exposed
fines a set of messages to be exchanged with a interfaces above and below the IUT. It may not be
peer entity; they are called Protocol Data Units applicable when conformance evaluation is done
(PDUs). ( N ) - P D U s are exchanged as data in the by a test laboratory because these of assumptions.
parameters of some of the ( N b - 1)-ASPs; other
( N b - 1)-ASPs convey control information to use 3.2. Distributed Method
the services of the ( N - 1)-layer. The test method
includes two logically distinct elements which are When an implementor (client) arranges for a
called the upper tester and lower tester because of test laboratory to test a protocol implementation,
their relationship to the interfaces of the IUT. direct access to the b o t t o m interface of the I U T is
In earlier work, an implementation of a lower not likely to be available (except in the data link
tester was called either a test driver or an en- layer where media standards require an exposed
coder/decoder; an upper tester was called a test interface). The distributed method (Fig. 2) is one
responder. But, ISO did not adopt these terms. of three models which make no assumptions about
The upper tester is assumed to generate and re- the existence of a PCO at the b o t t o m of the IUT.
R.J. Linn, Jr. / Conformance testing for 0.9 protocols 201

(4) synchronization and control are more dif-


ficult because elements of the test system are
distributed over two systems,
Synchronization and control (test coordination
procedures) may be specified by ASPS exchanged
at PCOs (or possibly by a test management proto-
col which is not standardized). The distributed
method relies on the protocol being tested to
provide sufficient synchronization to achieve test
purposes. Therefore, judgments and verdicts for-
mulated depend upon behavior observed by the
c I lower tester. In practice, there must be some coor-
dination between upper and lower testers; the
(N-l) Service Provider
same tests must be selected and executed concur-
Fig. 2. Distributed method.
rently. However, the distributed method does not
address how these issues are resolved.
In summary, the distributed method is a logical
equivalent to the local method with lower tester
and IUT interconnected by a communications
In all methods employing a communications service. However, the structure of the local method
service, IS0 denotes the complementary set of implicitly gives the capability to synchronize and
abstract service primitives available to the lower control the upper and lower testers because they
tester as (Nb - l)-ASP”s (double prime). are elements of the same test harness. Information
Note, the lower tester and IUT reside in two collection and sharing is possible (although a local
different systems. The lower tester and IUT are issue). Thus, the distributed method is not a func-
connected by an underlying OSI service which tional equivalent to the local method.
offers an (N - 1)-service using lower-layer proto-
cols and the physical media connecting the sys- 3.3. Coordinated Method
tems. Despite its name, the lower tester is obvi-
ously a peer entity of the IUT in this figure. The Two features that distinguish the coordinated
arrows between the IUT and OS1 service do not method (Fig. 3) from the distributed method are:
imply a real interface, just the conceptual flow of (1) no exposed upper interface is necessary
N-PDUs. Also note, an upper tester is part of the within the IUT (although this is not precluded);
model and an exposed top interface is assumed at and
the PCO.
Several consequences flow from changed as-
sumptions:
(1) abstract test cases written for the local
method are not applicable-they must be rewrit-
ten to reflect the abstract service primitives availa-
ble to the lower interface of the lower tester.
Generally, they are complementary to those as-
sumed available at the bottom interface of the
IUT in the local method;
(2) the lower tester and IUT are physically
separated with the implication that they observe
the same test event at different times;
(3) data loss, delivery out of sequence, and
data corruption are possible, particularly at lower
layers due to the quality of some lower-layer
services; Fig. 3. Coordinated method.
208 R.J. Linn, Jr. / Conformance testing for OSl protocols

(2) a standardized test management protocol ('3


Test Coordination
(TMP) and test management protocol data units procedures
(TMPDUs) are used to automate test management
and coordination procedures. Often it it is as-
sumed that the lower tester is the master and the
upper tester is a slave to minimize the effort in
Lower
realizing an upper tester. Tester ~ ~!~i~ Implementation
This is the most sophisticated model. It allows UnderTest
a very high degree of coordination and reporting
of information observed and collected at both the
upper and lower testers. Upper and lower testers (Nb-])-ASP"s
may be synchronized, information received by the t
upper tester may be reported back to the lower (N-l) Service Provider
tester for formulation of verdicts, selection of test Fig. 4. Remote method.
cases by an operator may be conveyed from the
lower tester to the upper tester, and branching
logic may be implemented for selecting tests if a communications below the network layer and
previous test case fails. To date, ISO has assumed above the transport layer raises serious prob-
minimal coordination (e.g., test cases will be ex- lems with exposed interfaces. Historically, it is
ecuted in a predefined order with no branching if generally accepted that the transport layer will
the verdict of a test case is fail). This assumption have an exposed interface and be the basis of a
can lead to serious synchronization problems if "test platform". Thus, the transport layer is a
the test management protocol does not resynchro- logical candidate for providing out-of-band
nize the upper and lower testers after a test fails. communications for upper-layer testing.
Communications between upper and lower tes-
ters may be in-band (TMPDUs are carried as data 3.4. Remote Method
via the protocol being tested) or out-of-band (use
of a lower-layer protocol which is assumed to be The last method defined by ISO for single layer
reliable enough to carry TMPDUs). Test manage- testing is the remote method (Fig. 4). The signifi-
ment using out-of-band services raises issues about cant features are that no interface at the top of the
which layer services are assumed to be exposed. I U T is assumed, and no explicit test coordination
Unfortunately, ISO conformance methodology procedures are assumed. (Test coordination, if any,
has not addressed the following issues: is manual.) The method relies solely on the proto-
- D e f i n i t i o n of a test-management-protocol col being tested for synchronization of the lower
kernel which is independent of its application. tester and the IUT. The method assumes that the
Without a standardized test-management-pro- state of the I U T is known from actions specified
tocol kernel, standards groups will have to in- for the lower tester including knowledge of N-
vest the effort to invent their own. This has PDUs transmitted and received via ( N b ) - A S P " s .
happened with the Session and Transport ab- Verdicts must be formulated based upon stimulus
stract test suites, and the test management pro- provided by the lower tester and the responses of
tocols are distinctly different. If test laborato- the IUT as observed by the lower tester. This
ries invented their own test-management proto- method is widely used for testing implementations
cols, test services ~ould be incompatible which of X.25.
is even worse. Currently, ISO is studying the
issue of a standardized kernel of test manage-
ment functions. 4. Tree and Tabular Combined Notation (TYCN)
- Recommendations for in-band or out-of-band
communications. Both are feasible; both have The usual assumption regarding conformance
drawbacks. In-band communications may be evaluation is: a test case (also called a test scenario)
difficult at the application layer but is feasible is written with a particular test purpose in mind.
in some cases (e.g., electronic mail). Out-of-band If test cases are to be the basis for I S O / O S I
R.J. Linn, Jr. / Conformance testingfor OSI protocols 209

conformance testing, it is clear that a notation or namic aspects of a test case, and additional tables
language is required to specify behaviors of a test to describe the static aspects of a test suite. For
system and a protocol entity to be tested. The example, tables are used to declare variables, iden-
language must contain sufficient features to de- tify points of control and observation, define ref-
scribe tests for all OSI protocols. With these pur- erences to other standards where data types are
poses in mind, ISO has defined a notation which defined (structure and fields of protocol data
is called Tree and Tabular Combined Notation units), and to specify assumed constraints. (This
(TTCN). It is defined in the third document (called form is often called the graphics form, or TI'CN-
Part 3 for brevity) of the ISO Conformance Test- GR.) For every tabular element in TTCN-GR,
ing Methodology and Framework [12]. A test case Part 3 also includes BNF syntax definitions for
written in TTCN specifies, as a sequence of atomic the machine processable form of TTCN (TFCN-
test events, the behaviors of a test system and a MP). TTCN-MP is also called the transfer syntax
protocol entity in sufficient detail to judge the of TTCN because it allows for electronic exchange
behavior of an implementation of the protocol to of test suites.
formulate a verdict of pass/fail. The dynamic elements of TTCN include assign-
TTCN is intended to be applied by protocol ment and arithmetic operators, predefined func-
developers within CCITT and ISO. They will de- tions (e.g., string manipulation), label declara-
fine abstract test suites. (An abstract test suite is tions, a goto statement (used to specify cyclic
independent of the features of a protocol incorpo- behaviors), and input/output operations. The send
rated in a particular IUT, and it is independent of operation is denoted "!"; receive is denoted as
any test system(s) employed by test laboratories to "?". Both may be qualified by the name of a point
conduct protocol testing. As such, abstract test of control and observation. These dynamic aspects
suites are not necessarily executable and TTCN of TTCN are presented on paper in tabular form.
was not intended to be an executable test lan- The column containing the text describing actions
guage.) Thus, when developing an abstract test is significant; i.e., blanks and tabs are significant
suite, a standards body may assume a particular in the semantics of the language. When describing
test method from those outlined in the previous a tree of possible input/output sequences, tabs are
section. However, standards bodies are not re- used as a shorthand notation for the text in the
sponsible for executable test suites or the imple- same column on previous lines; i.e., tabs are inter-
mentation aspects of the methods employed to preted as if the text above was repeated.
describe tests. An example may help. Consider a connection
To understand TTCN (and its name) one must oriented protocol which a lower tester employs as
understand the structure of the notation and what an ( N - 1) service and assume the distributed test
it is intended to express: TTCN is employed to method (Fig. 2). Five dynamic behaviors that
describe the dynamic behaviors and exchange of might be found in an input/output trace (log) are
messages between a test system and a protocol tabulated below. They are presented in Table 1 as
entity. A graph of the dynamic behavior of a sets of events and in Table 2 as a tree of behaviors
protocol entity may be drawn with the nodes of that may be observed at the PCO of the lower
the graph labeled with inputs and edges labeled tester (as they might be specified in TTCN).
with outputs (or vice versa). The resulting graph is (Nb - 1)-ASP's initiated by the lower tester are
a tree reflecting the dynamic behavior of a proto-
col entity in response to its inputs. (Similar infor-
Table 1
mation is often expressed as state graphs with
A set of dynamic behaviors for a connection oriented service
both inputs and outputs used to label the edges.)
Thus, TTCN is a language for describing trees of 1 Con.Req Con.Conf Dat.Req Dat.Ind Dis.Req
2 Con.Req Con.Conf Dat.Req Dat.Ind Dis.Ind
behavior and actions associated with sending/
3 Con.Req Con.Conf Dat.Req Dis.Ind
receiving test events and protocol data units 4 Con.Req Con.Conf Dis.Ind.
(PDUs). 5 Con.Req Dis.Ind
TTCN's name is derived from the fact that the Abbreviations: Con Connect Req Request
graphic definitions in Part 3 use trees of behavior Dat Data Con Confirm
specified in a tabular form to describe the dy- Dis Disconnect Ind Indication
210 R.J. Linn, Jr. / Conformancetestingfor OSl protocols

Table 2 named preamble, body, and postamble; each may


A tree of dynamic behaviors observable by a lower tester be composed of one or more named elements.
1 !Con.Req ?Con.Conf !Dat.Req ?Dat.Ind !Dis.Req (However, a p r e a m b l e and p o s t a m b l e are
2 ?Dis.Ind optional.) Since all identifiers are global, individ-
3 ?Dis.Ind ual elements of a test case m a y make reference to
4 ?Dis.Ind
previously identified objects (e.g., types, variables,
5 ?Dis.Ind
labels associated with a tree of behaviors). T T C N
-TIME allows reference to d e m e n t s of a test suite by
Legend: ! Send
? Receive attaching a sequence of named test steps. Concep-
tually, this is equivalent to the "include" or " c o p y "
facilities of some programming languages; how-
in bold font in both tables; those observed by the ever text substitution follows specific recursive
lower tester in response to actions taken by the rules. N a m e s m a y be qualified by other names
I U T are in R o m a n font. The numbers in the first denoting the inclusion of hierarchically structured
column of the two tables identify the correspond- elements of text.
ing set of events. In Table 2, empty entries in the An organization strategy for test suites sug-
left-hand columns of a row are interpreted as if gested by ISO is:
the text above it in the same column were copied - basic interconnection tests which are intended
into the row. Table 2 is not a test case. In a test to establish that an I U T conforms sufficiently
case, the entries in row 1 of Table 2 could be put to justify further testing (or identify severe cases
into separate rows and each entry in the table of non-conformance);
could be followed by additional test steps which - capability tests which are intended to establish
might include verdict assignment (verdicts are that the static conformance requirements of a
specified in a separate column). Comments are protocol are met (but not probe the I U T for
placed in the last column of a test case. A group detailed behaviors); and
of test steps may be named and referenced as behavior tests which are intended to test a full
-

group. range of dynamic behaviors which the supplier


In T T C N , the scope of variables is global. of a product claims to support.
Behavioral expressions are deterministic; i.e., text- Together, capability and behavior tests are in-
ual order of test events in a test case is used to tended to establish conformance or non-confor-
resolve nondeterminism. Basic assumptions un- mance of an IUT. Rayner [42] describes the pur-
derlying T T C N are that a protocol entity in an pose of this organization of tests in detail.
I U T can be driven into an assumed " s t a t e " b y a Operational semantics are being defined for
sequence of inputs and outputs (test steps). Once T T C N . (Early versions had no formal semantics.)
reaching that state, another sequence of test steps Without rigorous semantics:
judges the behavior of a particular aspect of the (1) T T C N is subject to subtly different inter-
I U T and a verdict is formed. Finally, the I U T is pretations; and
driven back to a known initial state by subsequent (2) it is possible for either humans or auto-
actions specified in a test case. This suggests a mated tools to interpret T T C N differently and
preamble, followed by a sequence of test steps, introduce errors when translating T T C N into an
followed by a postamble. The notions of preamble executable test suite.
and postamble are used in organization of a test Wiles [46] describes an environment for T T C N
suite where certain ~aamed groups of test steps which includes both a syntax directed editor and
may be used repeatedly (e.g., connection establish- interpretive execution. Currently, this is the only
ment and termination). This functional grouping environment known which directly executes ab-
of test steps leads to the notion of libraries of stract test cases. Probert et al. [39] are developing
named entries composing a test suite. an integrated environment for specification of
T T C N also facilitates organization of the ele- T T C N test suites which includes the facilities to
ments of a test into larger units and, ultimately, edit and transform canonical representations of
into a test suite; i.e., a hierarchical collection of test cases into T T C N - G R , T T C N - M P and an
test cases. A test case m a y be composed of a executable form. Others are also working on T T C N
R.J. Linn, Jr. / Conformance testingfor OSI protocols 211

tool kits. Since machine translation and execu- However, a non-standard T M P is unlikely to be
tion/interpretation of T T C N test suites is feasi- used in a standardized test suite.
ble, it is obvious that rigorous semantics are re-
quired if the intent of those individuals writing 5.1• The Testing Process
test suites are to have the same interpretations.
Thus far, the process of testing an I U T has
been ignored; ISO calls this a test campaign• Fig-
ure 5 depicts how testing proceeds given some of
5. Synthesis of the Concepts Presented
the elements discussed. It also introduces some
new elements. The following description is ideal-
Before introducing new topics, let us sum- ized: some steps m a y be combined, omitted
marize the discussion thus far. ISO has defined a n d / o r reordered.
T T C N as a language for the specification of ab- The process model assumes a standards com-
stract test suites for OSI protocols a n d I S D N mittee has done the following for each protocol to
systems. T T C N was not intended to be an execu- be tested:
table language. Nonetheless, tools are being built - developed at least one standardized abstract
to directly interpret or translate T T C N into execu- test suite for the protocol using one of the test
table tests. The g r a m m a r of T T C N - M P and oper- methods described earlier;
ational semantics provide necessary definitions to - developed a Protocol Implementation Confor-
support these developments. Thus, T T C N must be mance Statement (PICS) which defines a set of
considered an executable language. questions that a product supplier answers be-
ISO has defined four single layer test methods fore taking a product to a laboratory for test-
for OSI protocols. U p p e r and lower testers are ing;
abstractions of elements that m a y be found in real - developed a Protocol Implementation eXtra In-
test systems; they serve as virtual interpreters of formation for Testing (PIXIT) statement. The
T T C N within the constraints of a particular test P I X I T identifies implementation-specific infor-
method• mation regarding the I U T that is necessary for
These aspects are c o m m o n to the four test testing (e.g., timer values required, number of
methods: P D U s acknowledged in a single acknowledg-
- The role of the lower tester is to interpret ment, addressable elements in a stack of proto-
abstract test cases and formulate verdicts based col entities).
upon behaviors specified in a test case. (Note, The ISO test methodology identifies the latter two
there is no requirement that elements of a real items as required parts of a protocol standard. At
test system formulate verdicts in real-time.)
- The lower tester acts as a peer entity of the I U T
and exchanges N - P D U s with the IUT.
I of Static Analysis
Options
I
These are the differences in the test methods:
- Control and observation by the lower tester is
specified in terms of:
- ( N b - 1)-ASPs in the local method;
No > Verdict: Fail
- ( N b - 1 ) - A S P ' s in the other methods.
- The remote method assumes no control and
observation using an upper tester.
- The coordinated method assumes no upper in- Test Parameterization I
terface and employs T M P D U s to control the
upper tester. Test Execution ~ SCTR~
- The local and distributed methods assume an
upper tester plus control and observation Results Analysis

specified in terms of (Nt)-ASPs.


- The distributed method m a y employ a non-
standard test management protocol (TMP). Fig. 5. Test campaign.
212 R.J. Linn, Jr. / Conformance testingfor OSI protocols

this time, these elements do not exist for all OS1 tocol layer at a time within an IUT which is
protocols. embedded one or more layers down in a stack of
A test laboratory must either have the means to protocols within an IUT. Usually, testing proceeds
interpret an abstract test suite directly or have bottom-up within the stack, and upper layers are
transformed it into an executable test suite by tested incrementally as conformance of lowers is
some means. If a laboratory offers more than one established; i.e., focus of testing switches between
test method, it is the client’s choice which will be layers as testing proceeds. IS0 has defined em-
used to test an IUT. The information contained in bedded testing for each of the four methods out-
PICS provided by a client identifies what options lined in Section 3.
and features of a protocol have been implemented. Conceptually, it is easy to consider lower layer
The client’s PICS is used by a test laboratory for protocols as envelopes wrapped around upper
static conformance assessment. Specifically, a layer PDUs. This is certainly true for the data
laboratory checks to assure that mandatory fea- transfer phase. At each successive layer, a header
tures are included and conditional features (those prefixes “data” from the adjacent upper layer. It
dependent upon other options) are also imple- is this relatively simple model that underlies the
mented as required. The PICS is also used for test notions of embedded testing. Assume that the
selection; i.e., the laboratory eliminates tests for N-layer is to be tested and (N + l)-PDUs are to
features and options which are not implemented. be embedded in N-PDUs. Conceptually, two-layer
Information contained in the PIXIT is used to embedded testing defines a set of (N + l)-PDUs
parameterize a test suite. For example, timers and and then envelopes this set of data in appropriate
addresses of each component protocol of the IUT N-PDUs. This model is depicted in Fig. 6.
may have to be set before testing commences. The direct consequence of these assumptions is
Tests are executed, results are analyzed, and that each test case must concurrently specify all
two reports are generated. A Protocol Confor- the actions of two layers (or at least a functional
mance Test Report (PCTR) contains sufficient subset of the (N + 1)-layer) in order to achieve a
information to uniquely identify the system and test purpose. A test case must have a specific test
protocol tested, identifies the test suite and test purpose and must reflect dynamic behaviors possi-
method employed, and contains verdicts for each ble for two layers of protocol in the IUT. Ad-
test case executed (or a note indicating that a test ditionally, test cases must reflect the context of an
was not executed) and identifies the conformance (N + l)-protocol in the lower tester (e.g., results of
log supporting verdicts rendered. Since more than negotiating options, or variables which record pro-
one protocol may be tested, there may be more tocol control information that influence subse-
than one PCTR. A System Conformance Test quent behavior) as the protocol progresses through
Report (SCTR) identifies a system which was its phases (e.g., connection establishment, data
tested and gives a summary conformance state- transfer, and connection termination). A test suite
ment for each component protocol tested.

Upper
6. Embedded and Multi-layer Testing Tester

Thus far, single layer test methods were pre-


sented. In the following subsections, we look at
ISO’s embedded test methodology, and then re-
port the status of multi-layer testing within ISO. Under Test

6.1. Embedded Methodr


1 1 (Nb-I)-ASP?
Currently, IS0 has only defined what are known
I t I
as the embedded methods for testing one or more
layers of a multi-layer IUT without exposed layer I ’ (N-l)-Service

interfaces. Embedded testing focuses on one pro- Fig. 6. Embedded method with test coordination.
R.J. I.a'nn,Jr. / Conformance testingfor OSI protocols 213

consists of describing all possible behaviors be- the File Transfer and Management, and Message
tween two peer protocol entities and reflecting the Handling Systems protocols).
context of the (N + 1)-protocol in order to pre-
vent a test from being aborted. In fact, this be- 6.2. Multi-Layer Testing
comes quite complex very rapidly and becomes
progressively more difficult when attempting to
Part I of the ISO documents defines multi-layer
specify behaviors for more than two layers. At this
testing as "Testing the behavior of a multi-layer
time, only limited experience exists with the
IUT as a whole, rather than testing it layer by
method. It is viable for two layers, but usually,
layer". However, no further guidance is given on
test cases use a relatively small subset of the
the topic and currently multi-layer testing is the
(N + 1)-protocol.
subject of research.
The advantage of an embedded method is that
Advocates of the Ferry Clip method (intro-
it does not assume an exposed interface at every
duced in the next section) suggest that it may be
layer of the IUT. The disadvantages are:
used for multi-layer testing. However, an assump-
- test cases become more complex in a non-linear
tion they make is that the IUT has exposed inter-
relationship for each additional layer above the
faces at each layer which is not valid for many
layer to be tested;
commercial products. An example of one formal
- it is correspondingly more difficult to anticipate
method to conduct multi-layer testing is included
all possible behaviors given that suppliers have
in the next section. While the methodology em-
valid implementation options and choices for
ploys Estelle and ASN.1, the methodology may
each protocol that is involved in the test. Thus,
also be exploited by proponents of LOTOS and
verdict assignment is more difficult and the
SDL.
probability of either an inconclusive or invalid
verdict increases;
- test suites are defined assuming a particular set
of protocols; a new test suite must be written to 7. O p e n Topics within I S O Conformance Method-
test the same layer if the combination of proto- ology
cols above N-layer change. Furthermore, if any
protocol in the combination changes, the entire While the following topics are under study by
test suite must at least be checked, if not en- ISO, we include them as part of an overview of
tirely rewritten. conformance testing for OSI. The first is ferry test
Due to the reasons identified above, embedded methods. Some argue that the ferry methods are
testing presents some significant problems to those .simply a means to realize the test coordination
defining abstract test suites for embedded meth- implied by either the distributed or coordinated
ods. Test laboratories employing embedded meth- methods. Others believe they are distinct test
ods face additional problems: they must obtain or methods. Regardless of opinion, ISO has been
create, and subsequently maintain, executable test requested to consider the ferry methods as one
suites and systems which reflect the intent of the possibility of realizing test coordination as part of
abstract test suites. a larger question: definition of a kernel set of test
Given the possibility of single-layer and em- management functions a n d / o r test management
bedded methods, ISO could produce test suites for protocol.
every possible combination of local, distributed, Second, ISO is studying the potential roles of
coordinated and remote methods and groupings of formal methods and formal description techniques
protocols. This would require significant resources within conformance testing. The scope of this
and add to the complexity of testing. In most topic includes generation of test sequences as well
cases, standardized abstract test suites are likely to as application of the formal description techniques
be written employing one method per protocol (or named Estelle [9], LOTOS [10] and SDL [18].
group of protocols). But, there are exceptions: Since an explanation of test generation method-
abstract test suites being developed by ISO for the ology is beyond scope of this paper and the topic
Session protocol include the single-layer coordi- is surveyed in another paper [38], only some of the
nated and single-layer embedded methods (under most recent research results are noted below.
214 R.J. Linn, Jr. / Conformance testing for OSI protocols

Sabnani and Dahbura, and Aho et al. [43,23] (although not explicitly identified). It is called an
describe methods to optimize generation of test "active" ferry entity because it assumes the role of
sequences. A proposal was made to incorporate a master; the entity above the IUT is called a
results of optimization in the developing abstract "passive" entity because it responds to actions
test suite for X.25. Sidhu and Leung [44] report on taken by and directives issued by its peer. The
the ability of several test generation methods to following summarizes the concepts found in the
detect faults. Ural [45] defined a method for selec- original work and assumes in-band communica-
tion of test cases based upon static control flow tions except as noted.
and data flow analysis of descriptions of protocols The ferry control entity serves as a carrier of
written in Estelle. Barbeau and Sarikaya have PDUs received from either the upper or lower
developed a computer-aided design tool [24] which testers by retransmitting them to the other entity;
can display both control flow and data flow graphs i.e., it functions in logical loop back mode.
of protocol specifications written in Estelle. For- Minimal state information and headers are re-
ghani and Sarikaya have created a tool to generate quired to realize a ferry control implementation,
TTCN test suites from EsteUe specifications [31]. and it can be designed in a relatively layer-inde-
To date, most applications of LOTOS have focused pendent manner. The advantages of the ferry con-
on protocol specification and verification rather trol protocol are:
than testing. Exceptions are the work of Brinksma (1) it places a relatively small burden on an
[27] and de Meer [30]. implementor,
The following subsections report on topics (2) it is protocol independent,
raised in ballots on the ISO documents and are (3) it is independent of test management proto-
outstanding questions to be resolved. col, and
(4) most test coordination problems are the
7.1. Ferry Control and Ferry Clip Methods test laboratory's.
Its disadvantages are:
Zeng [47] defined an alternative to both the (1) an exposed interface is assumed at each
distributed and coordinated methods. The upper layer of the IUT,
tester is moved from the client's system to the test (2) synchronization problems unique to the
laboratory's system. The upper tester in the client's ferry method may arise (it is a state based proto-
system is replaced by a ferry control protocol col entity),
entity (the name was originally derived from the (3) it requires an "interface adaptor" for each
notion of a ferry boat). The model is depicted in layer, and
Fig. 7. Note, the block labeled "Test Coordination (4) it simply may not be applicable for Appli-
Procedures" includes a peer ferry control entity cation and Presentation layer protocols.
The ferry method does not resolve the in-band/
out-of-band communications issues described
Ferry earlier. Either may be used if the state of the IUT
Upper Control is not changed by conveying data to and from the
Tester Protocol
upper and lower testers; otherwise, out-of-band
I ; | communications is required. In his original work,
Test (Nt)-ASPs
Coordination: Zeng proposed out-of-band communications via a
I Procedures I I so-called "ferry control channel". In Fig. 7, some
Implementation
reliable (N - k)-service may be used to realize the
Lower
Tester N-PDUs~ Under Test ferry control channel (e.g., Transport protocol) if
there is an exposed lower layer interface. Other-
wise, the ferry control channel may have to be
(Nb-1)-ASP"s derived by some other medium.
p,
Extensions Zeng has proposed to his original
work are known as the ferry clip method [48]. In
(N-l) Service Provider
brief, the ferry clip method assumes that exposed
Fig. 7. Ferry method. interfaces exist at all layers of interest in the IUT,
R.J. Linn, Jr. / Conformance testingfor OSI protocols 215

and that the passive ferry entity can be attached to


Lower Upper
multiple PCOs in the IUT. Thus, the ferry entities Tester Test Mgmt Proel Tester
must multiplex streams of data t o / f r o m upper TMPDUs
and lower testers for multiple layers of protocol -

within the IUT. Additionally, the ferry entities


must provide a reliable data transport mechanism Layer N+I •~ (N+I)-PDUs"~
Irnplementatior
when testing lower layer protocols (e.g., network UnderTest
and below) if the method is to provide a reliable
means of observation and control. It remains to be Layer N ~ N-PDUs
seen if ISO will adopt either ferry method.

7. 2. Architectural Refinement "~(Nb-1 )-ASP"s

Current ISO methodology restricts the number (N-1)-Service


of points of control and observation in a lower Fig. 8. Architectural refinement.
tester to exactly one PCO. Therefore, it is impossi-
ble to make reference to any PCO other than the
one immediately above the ( N - 1 ) service.
Specifically, within a test case specified in TTCN, protocol context for one or more layers above the
it is impossible to make reference to (Nt)-ASP"s layer under test (Fig. 7). Indeed, Davis [29] de-
containing (N + 1)-PDUs .which might be gener- scribes the architecture of a test system which
ated by any other higher layer entity within a employs reference entities above and below the
lower tester; i.e., a "higher-layer test case" or layer under test. Current practice at the Corpora-
reference entity. For our purposes, assume a refer- tion for Open Systems (COS) demonstrates the
ence entity is: viability of the concept. COS employs a reference
(1) an implementation of a protocol which sup- implementation of the Transport Class 4 protocol
ports all functions and options specified in a pro- above the test entity for the CLNP (Protocol for
tocol standard, Providing the Connectionless Network Service).
(2) can interoperate with any implementation Reasonable arguments support developing a
which realizes a valid subset of behaviors specified standardized test method which allows reference
in a standard, and entities or single-layer test entities to be inter-
(3) has been tested sufficiently to demonstrate changed at every layer of a real test system. For
the first two features. example, ISO's FTAM and CCITT's MHS proto-
While a reference entity may have been derived cols use different functional subsets of the Session
from a formal description of a protocol by auto- protocol and may run over different Classes of the
mated methods, this is not strictly necessary. Transport protocol. Maintenance of test suites for
Whether or not an IUT is implemented with embedded methods will be expensive to ISO and
interfaces at each layer, a test method may be test laboratories. Perhaps, the strongest argument
developed with layered components, each focusing is ISO's definition of multi-layer testing: "Testing
on a specific aspect of testing a multilayer imple- the IUT as a whole".
mentation. Such a test method (system) acts as if Given the approach depicted in Fig. 8 and a
the IUT is layered, but it does not depend upon multi-layer test method, the N-layer could be
the internal structure of the IUT as no assump- tested with an arbitrary combination of protocols
tions regarding layering and interfaces are neces- above it. Note, this approach de-emphasizes the
sary. Thus, the OSI Reference Model [6] suggests role of test cases and puts the emphasis on refer-
an alternative to embedded methods. The Refer- ence entities. The consequences are:
ence Model refines the problem of communica- (1) for purposes of testing, the test case/behav-
tions into layers of protocols, each with distinct ior observed focuses on a single layer; and
functions and services. This approach suggests (2) test scenarios play a significantly smaller
reference entities could also be used above the role and consequently become significantly sim-
N-layer in multi-layer test systems to maintain the pler.
216 R.J. Linn, Jr. / Conformancetestingfor OSl protocols

This architectural refinement of a "lower tester" changes to a formal description of a standardized


into a stack of reference entities above and below protocol:
the N-layer raises new issues: (1) generation and transmission of valid PDUs
- What role do test scenarios play and how do at points disallowed in the standard (inopportune
they differ from single layer test scenarios? PDUs);
- What points of control and observation: (2) transmission of parameters of PDUs which
- are relevant within a test system; and are either out of range or semantically inconsistent
- how is error recovery on the part of the IUT with the protocol standard; and
tested? (3) invalid encoding of PDUs.
- If any protocol entity above or below the layer The first two cover most dynamic aspects of
under test behaves invalidly: testing error recovery mechanism in a standard.
- how can this be observed; The first can readily be achieved by defining ac-
- what verdict should be assigned; tions in response to ( N t ) - A S P " s that are never
- what information should be logged; and valid user sequences (e.g., request data transmis-
- how should reference entities be specified sion before a connection is established). The sec-
a n d / o r realized? ond set can be achieved by defining additional test
If it is desirable to observe and report invalid services which override normal protocol con-
behavior observed on the part of the I U T in more straints; i.e., define specific actions in response to
than one layer, nothing prohibits a test system new abstract service primitives. Note, these ad-
designer from extending a reference entity into a ditional ( N t ) - A S P " s would not be invoked by a
test entity with this feature. "normal user" and therefore will have no effect
unless specifically invoked in either a single-layer
or multi-layer test system. The third set raises
pragmatic issues. Totally undecodable PDUs are
7.3. Formal Specification of Multi-layer Test Sys-
likely to generate inconclusive results unless the
tems
protocol has specific error recovery mechanisms
for loss and corruption of PDUs. Exhaustive test-
CCITT has produced a Formal Description ing in this domain is infeasible. Thus, the usual
Technique (FDT) named System Design Lan- approach is to intuitively construct a small set of
guage (SDL) [18]. Jointly, ISO and CCITT have tests covering likely failures on the part of imple-
produced Abstract Syntax Notation One (ASN.1) mentors.
[7-8]. ISO has produced two FDTs, Estelle The approach outlined above is more than the-
[9,28,35] and LOTOS [10,26]. They became inter- ory. Figure 9 depicts the refined architecture of a
national standards in 1988; thus, their application test system developed at NIST for testing FTAM
has been limited. To date, most OSI protocol through the Presentation layer protocols using
standards have used natural language and state either the remote or coordinated methods. Thir-
tables to define the standard. ASN.1 has been teen modules are specified in Estelle: the Lower
employed to define message syntax for upper layer Tester Manager down through a generic Session
protocols. If standardized FDTs are applied to interface. There are two major clusters of mod-
protocol specification, then it is quite natural to ules:
augment their formal descriptions using the same (1) the upper cluster of modules which were
F D T to specify the behavior of a test entity. Since designed to test F T A M [14-17] and the layers
translators for the FDTs exist, it is also quite below it, and
natural to realize executable test entities by trans- (2) the lower cluster of modules which are aug-
lation. mented test entities that implement the FTAM,
However, testing requires mechanisms to inject ACSE and Presentation protocols. All are derived
errors to test for error recovery on the part of the from Estelle specifications of the corresponding
IUT. A logical way to proceed is to simply extend protocol entities. Estelle is used to specify and
a standardized protocol specification by specifying define the dynamic aspects of the upper cluster of
the behaviors necessary to invoke error recovery modules and the stack of protocols in the lower
on the part of the IUT. This implies three basic cluster. Augmented ASN.1 type definitions are
R.J. La'nn,Jr. / Conformance testingfor OSI protocols 217

LOWER TESTER ject of open debate. Fault detection is not only


related to the test method, but is related to fault
models, the test language employed, test suite
coverage, and synchronization issues. If all were
equal (and they are not), assumptions often made
are: from most powerful to least, the fault detec-
tion capacity of ISO's test methods is local, coor-
dinated, distributed and remote methods for single
layer testing. Embedded methods limit observabil-
ity and control and are generally considered
weaker than single layer methods. Proponents of
the ferry methods argue that ferry methods have
equal fault detection capacity as the local method.
The Conformance Testing Service (CTS) pro-
jects (CST-WAN and CTS-LAN), sponsored by
the Community of European Countries (CEC),
have produced a substantial number of test tools
and test suites. Many will be employed within
European test centers which have been established
to test OSI products for use throughout the CEC.
Information regarding the tools and test suites
may be ordered from the National Computing
Centre in the UK.
Fig. 9. Multi-layertest architecturefor the FTAM,ACSEand
presentation protocols. Parts 1, 2, 4, and 5 of the ISO testing methodol-
ogy should advance to international standard
used to specify the encoder/decoders for the pro- status within a year; Part 3 is likely lag by up to
tocol stack and to define the test language of the 18 months. CCITT is applying TTCN to standard-
system. ized abstract test suites for the X.400 series recom-
Regardless of the internal structer of an IUT, mendations for electronic mail (Message Handling
the components of the test system are able to Systems), X.25 and its components, and ISDN
observe behavior at all three layers, detect and D-channel testing. These are destined to become
report invalid behavior on the part of the IUT, CCITT recommendations by the end of the cur-
and log observed behavior (as ASN.1 values). rent study period (1992). ISO has started develop-
These components are also employed as part of a ment of abstract test suites for Transport, Session
system designed to test an application layer gate- and upperlayer protocols in TTCN. If current
way for ISO's FTAM and DoD's File Transfer schedules within ISO are met, many of these test
Protocol (FTP). Translators for Estelle [22] and suites will become international standards be-
ASN.1 [32,33] are used to realize implementations tween 1990 and 1992.
of both test systems. Another test system for an Current testing practice is best described as a
electronic mail gateway employs the same meth- patchwork quilt of old and new test systems (with
odology. Details of these systems [20,21] and the quite a few holes). The Corporation for Open
methods employed to realize them are reported Systems (USA) and National Computing Centre
elsewhere [36-38]. (UK) are using a mixture of older tools (adapta-
tions of 1985 and earlier vintage tools) to test
lower-layer protocols and newer tools developed
8. Concluding Remarks by the National Computing Centre (UK) for File
Transfer and Management (FTAM) and Message
We conclude with comments on fault detection, Handling Systems (MHS). The gaping holes are
tool development, progress of testing standards test suites.
and current testing practice. The fault detection Many test suites written in TTCN for MHS
capacity of the test methods described is the sub- and X.25 are available; FTAM test cases ~are
218 R.J. Linn, Jr. / Conformance testing for OSl protocols

scarce. Older tools use ad hoc test suites; but their Other Publications
history of use makes them relatively complete. [18] SDL, CCITT Recommendations Z.101-Z.104 (Blue Book
Series), Consultative Committee for International Tele-
graph and Telephone, International Telecommunications
Union, Place des Nations, CH 1211, Switzerland, 1988.
References [19] Message Handling Systems, CCITT Recommendations
X.400, X.401, X.408, X.409, X.410, X.420, X.430 (Red
Book Series), 1984.
IFIP: International Federation for Information Processing [20] A Test System for Implementation of M H S / S M T P Gate-
[1] Protocol Testing - Towards Proof, D. Rayner and R.W.S. ways, National Institute of Standards and Technology,
Hale, eds., Vol. 1-2, INWG/NPL Workshop, National National Computer and Systems Laboratory, ICST/SNA
Physical Laboratory, Teddington, Middlesex, T W l l OWL, 87/5, Parts 1-3, September 1987.
United Kingdom, 1981. [21] A Test System for Implementations of F T A M / F T P Gate-
[2] Proc. Protocol Specification, Testing, and Verification II, ways, National Institute of Standards and Technology,
C. Sunshine, ed. (North-Holland, Amsterdam, 1982). National Computer and Systems Laboratory, ICST/SNA
[3] Proc. Protocol Specification, Testing, and Verification III, 88/6, Parts 1-3, October 1988.
H. Rudin and C.H. West, eds. (North-Holland, Amster- [22] Users Guide for the NBS Prototype Compiler for Estelle,
dam, 1984). NBS, ICST, ICST/SNA 87/3, (Rev.) January 1989.
[4] Proc. Protocol Specification, Testing, and Verification IV, [23] A.V. Aho, A.T. Dahbura, D. Lee and M.U. Uyar, An
Y. Yemini, R. Strom and S. Yemini, eds. (North-Holland, Optimization Technique for Protocol Conformance Test
Amsterdam, 1985). Generation Based on UIO Sequences and Rural Chinese
[5] Proc. Protocol Specification, Testing, and Verification V, Postman Tours, in: Proc. Protocol Specification, Testing
M. Diaz, ed. (North-Holland, Amsterdam, 1986). and Verification VIII (North-Holland, Amsterdam, 1989).
[24] M. Barbeau and B. Sarikaya, A Computer-Aided Design
ISO: International Organization for Standardization, ISO Tool for Protocol Testing, in: Proc. INFOCOM '88 (IEEE,
Secretariat for JTC1/SC21, ANSI, 1430 Broadway, New York, New York, 1988) 86-95.
NY 10018, USA. [25] G. Bochmann, Usage of Protocol Development Tools, in:
[6] Information Processing Systems - Open Systems lntercon- Proc. Protocol Specification, Testing and Verification VI1
nection - Basic Reference Model, IS 7498, 1984. (North-Holland, Amsterdam, 1988) 139-161.
[7] Information Processing Systems - Open Systems Intercon- [26] T. Bolognesi and E. Brinksma, Introduction to the ISO
nection - Specification of Abstract Syntax Notation One Specification Language LOTOS, Comput. Networks ISDN
(ASN.1), IS 8824, 1987. Systems 14 (1) (1987) 25-60.
[8] Information Processing Systems - Open Systems Intercon- [27] E. Brinksma, A Theory for the Derivation of Tests, in:
nection - Specification of Basic Encoding Rules for Ab- Proc. Protocol Specification, Testing, and Verification VIII
stract Syntax Notation One (ASN.1), IS 8825.2, 1987. (North-Holland, Amsterdam, 1989).
[9] Estelle: A Formal Description Technique Based on an Ex- [28] S. Budkowski and P. Dembinski, An Introduction to
tended State Transition Model, IS 9074, 1988. Estelle: a Specification Language for Distributed Systems,
[10] Information Processing Systems - Open Systems lntercon- Comput. Networks ISDN Systems 14 (1) (1987) 3-24.
nection - LOTOS - A Formal Description Technique [29] W.B. Davis, Architecture and Design of a Portable OSI
Based on Temporal Ordering of Observed Behavior, IS Protocol Tester, in: Proc. 1st International Workshop on
8807, 1988. Protocol Testing, Comput. Sci. Dept., Univ. British Col-
[11] Information Processing Systems - 0S1 Conformance Test- umbia Vancouver, B.C., Canada V6T 1W5 (1988).
ing Methodology and Framework, ISO/IEC JCT 1/SC 21 [30] J. de Meer, Derivation and Validation of Test Scenarios
DIS 9646, Parts 1-2, November 1988. Based upon the Formal Specification Language LOTOS,
[12] Information Processing Systems - 05rI Conformance Test- in: Proc. Protocol Specification, Testing, and Verification
ing Methodology and Framework, ISO/IEC JCT 1/SC 21 111 (North-Holland, Amsterdam, 1987) 203-216.
N3077, Part 3, February 1989. [31] B. Forghani and B. Sarikaya, Automatic Dynamic Behav-
[13] Information Processing Systems - OSI Conformance Test- ior Generation in TTCN Format from Estelle Specifica-
ing Methodology and Framework, ISO/IEC JCT 1/SC 21 tions, Dept. of Electrical and Computer Eng., Concordia
DIS 9646, Parts 4-5, March 1989. Univ., 1455 de Marsionneuve Blvd. W. 915, H3G 1M8,
[14] Information Processing Systems - Open Systems Intercon- Canada, 1889.
nection - File Transfer, Access and Management Part I: [32] P. Gaudette, S. Trus and S. Collins, An Object-Oriented
General Introduction, IS 8571/1. Model for ASN.1, in: K. Turner, ed., Proc. FORTE'88
[15] Information Processing Systems - Open Systems Intercon- (North-Holland, Amsterdam, 1989) 121-134; also Report.
nection - File Transfer, Access and Management Part H: No. ICST/SNA 88/4.
The Virtual Filestore, IS 8571/2. [33] P. Gaudette, S. Trns and S. Collins, ASN.1 Free Value
[16] Information Processing Systems - Open Systems Intercon- Tool, National Bureau of Standards, Institute for Com-
nection - File Transfer, Access and Management Part IH: puter Sciences and Technology, ICST/SNA 88/2, January
File Service Definition, IS 8571/3. 1989.
[17] Information Processing Systems - Open Systems Intercon- [34] R.J. Linn, Testing to Assure Interworking of Implementa-
nection - File Transfer, Access and Management Part IV: tions of ISO/OSI Protocols, Comput. Networks 11 (4)
File Protocol Profile, IS 8571/4. (1986) 277-286.
R.J. Linn, Jr. / Conformance testing for OSl protocols 219

[35] R.J. Linn, The Features and Facilities of Estelle, National Tests, in: Proc. Protocol Specification, Testing, and Verifi-
Bureau of Standards, Institute for Computer Sciences and cation V (North-Holland, Amsterdam, 1986) 441-460.
Technology, ICST/SNA 87/6, November 1988 (Rev.) [42] D. Rayner, OSI Conference Testing, Comput. Networks
[36] R.J. Linn and J.P. Favreau, Application of Formal De- ISDN Systems 14 (1) (1987) 79-98.
scription Techniques to the Specification of Distributed [43] K.K. Sabnani and A. Dahbura, A Protocol Test Genera-
Test Systems, in: Proc. 1NFOCOM '88 (IEEE, New York, tion Procedure, Comput. Networks 1SDN Systems 15 (4)
1988) 96-109; also, Report No. ICST/SNA87/9. (1988) 285-297.
[37] R.J. Lima, J.P. Favreau, L. Gebase and A. Iwabuchi, An [44] D. Sidliu and T.K. Leang, Fault Coverage of Test Meth-
Overview of Formally Specified Multi-Layered Test Sys- ods, in: Proc. INFOCOM "88 (IEEE, New York, 1988)
tems, National Bureau of Standards, Institute for Com- 80-85.
puter Sciences and Technology, ICST/SNA 88/1, January [45] H. Ural, Test Sequence Selection Based on Static Data
1988. Flow Analysis, Comput. Comm. 10 (5) (1987) 234-242.
[38] R.J. Lima, Conformance Evaluation Methodology and [46] A. Wiles, B. Pehrson, ITEX-An Interactive TTCN Editor
Protocol Testing, IEEE J. Select. Areas Comm. 7 (7) and Executer, Swedish Institute of Computer Science,
(1989) 1143-1158. Design Methodology Laboratory, Box 1263, S-163 13
[39] R.L. Probert, H. Ural and M.W.A. Hornbeek, An In- Spanga, Sweden.
tegrated Software Environment for Developing & Validat- [47] H.X. Zeng and D. Rayner, The Impact of the Ferry
ing Standardized Conformance Tests, in: Proc. Protocol Concept on Protocol Testing, in: Proc. Protocol Specifica-
Specification, Testing, and Verification VIII (North-Hol- tion, Testing, and Verification V (North-Holland,
land, Amsterdam, 1989). Amsterdam, 1986).
[40] D. Rayner, Towards an Objective Understanding of Con- [48] H.X. Zeng, X.F. Du and C.S. He, Promoting the "Local"
formance, in: Proc. Protocol Specification, Testing, and Test Method with the New Concept "Ferry Clip", in:
Verification 111 (North-Holland, Amsterdam, 1984). Proc. Protocol Specification, Testin~ and Verification VIII
[41] D. Rayner, Towards Standardized OSI Conformance (North-Holland, Amsterdam, 1989).

Vous aimerez peut-être aussi