| Rfc | 1107 |
|---|
| Title | Plan for Internet directory services |
|---|
| Author | K.R. Sollins |
|---|
| Date | July 1989 |
|---|
| Format: | TXT, HTML |
|---|
| Status: | INFORMATIONAL |
|---|
|
Network Working Group K. Sollins
Request for Comments: 1107 M.I.T. Laboratory for Computer Science
July 1989
A Plan for Internet Directory Services
Table of Contents
1. Introduction 1
1.1. The Issues 1
1.2. Project Summary 3
2. Goals and Requirements for a White Pages Service 6
3. Pre-existing Services 9
4. Proposed Approach 11
4.1. Stage 1: The Field Test 12
4.2. Stage 2: Implementation 17
4.3. Stage 3: Deployment 17
5. Conclusion 18
Status of this Memo
This memo proposes a program to develop a directory service for the
Internet. It reports the results of a meeting held in February 1989,
which was convened to review requirements and options for such a
service. This proposal is offered for comment, and does not
represent a committed research activity of the Internet community.
Activity in this area is anticipated, and comments should be provided
promptly. Distribution of this memo is unlimited.
1. Introduction
1.1. The Issues
As part of the planned growth of the Internet (in particular, in
support of the full science research community in the U.S.), an
increasing need is anticipated for various sorts of directory
services. The increase in the size of the community served by the
Internet and the burgeoning demands for electronic mail lead to the
need for a service to find people's computer mailboxes and other
relevant facts, a so-called "White Pages" service. At the user level
to date, there have been no such national or international white
pages services in general use. As part of building the National
Research Network (NRN), it is important that such a service exist,
not only within the NRN community, but also crossing the boundaries
from the NRN to the more global network community. This will enhance
communication not only among computer scientists, but also among
scientists and engineers in other fields as well. Also important and
related is a so-called "Yellow Pages" service, which permits the
location of Internet resources based on their attributes.
A "White Pages" service is one in which one can look up people in
order to learn information about them for finding them. In its
simplest form, a white pages service provides what the white pages
telephone book provides. Based on a name, one can find an address
and a telephone number. In a network environment, there may be many
other kinds of location information, such as electronic mailbox,
electronic calendar, or file server, where one might leave a file for
the recipient. In addition, the electronic white pages may support a
much more sophisticated set of mechanisms for lookup. One might
match on a more complex set of attributes than first and last name.
In addition, the searching might span more than one local white pages
service. There are a number of naming and directory service
specifications and implementations in the field. They have differing
functionality and mechanisms to address that functionality.
Within the the world of networking today, there are a number of
partial solutions to the directory service problem. Examples of
these are the Internet Domain Naming Service (DNS), Clearinghouse,
DECnet Network Architecture Naming Service (DNANS), Profile, and
X.500. The Domain Naming Service provides a directory service most
commonly used for host naming and mail delivery. Clearinghouse and
DNANS are respectively the Xerox and DEC corporate naming services,
originally for mail delivery, although having other uses as well, in
both cases. Profile is part of the work of Larry Peterson to explore
descriptive naming in a non-hierarchical structure.
There is a CCITT recommendation X.500 (ISO DIS 9594), which defines a
general directory service. One of its primary goals is the naming
service needed for message handling (X.400). While X.500 is still
developing, and would need further evolution to cover all the
requirements of a service for the Internet, it will have an important
impact on the Internet community. It will form the basis of
commercial products, and it will almost certainly be the directory
service of many parts of the network world, which implies a need to
interoperate at a minimum. There is some concern that despite the
fact that X.500 is a recognized standard, there are a number of gaps
and limitations of the approach, that in turn will cause it to be
inadequate for the needs of the NRN.
In this context, a meeting was held to review current requirements
and solutions for directory services. This RFC reports the results
of that meeting, including the possibilities for a program of work in
this area.
For two days, a group representing academic, commercial, and
government interests in directory services discussed both alternative
candidates for a white pages service and the issues in building any
such service. The meeting was kept small by inviting only a small
number of representatives of each perspective. By the conclusion of
the second day, a consensus was reached on how one could achieve a
white pages service in three years. This is summarized in the next
section.
1.2. Project Summary
The consensus of the meeting can be summarized in the following five
points:
1. The standards and implementations are close enough to being
complete that it is reasonable to undertake provision of an NRN
"White Pages" service.
2. Although we are close, an effort is needed to experiment with
different levels of service, to flesh out the standards, and to
develop code.
3. An initial evaluation experiment is needed before making final
detailed plans for a production version of the service.
4. With strong funding and encouragement, a production service is
possible in three years.
5. It is important to act now to provide a coherent solution.
This means both having an impact on the evolving standards
and providing a unified, wide-spread solution before a plethora
of differing solutions appear.
Although it has clearcut drawbacks, X.500 was identified as the most
likely candidate directory service. The reasons for this are that it
has rich semantics and is becoming the accepted international
standard. However, there are problems with its incompleteness and
with its strict hierarchy. Therefore, in order to explore these and
become convinced of its viability, the consensus at the meeting was
to propose field trials, as the project's first stage. The field
trials would be limited in the user community, perhaps restricted to
computer science departments because of their familiarity with the
problems, and would be based on experimental or new software. They
would include experiments with at least an X.500 implementation,
Profile, and DNANS. Each of these services has strong points that
must be considered as part of the evaluation. They are:
X.500: International standard, hierarchy, search rules and
filters for searching attributed based names.
Profile: Descriptive naming with a richer semantics for
describing search criteria, an arbitrary network
of servers.
DNANS: Access control, replication, caching, hierarchy.
In summary, the plan would fall into three stages as follows:
- Stage 1: Field Trials.
There are two aspects to the field trials. The first is to
explore several different architectures for a white pages
service. To this end, implementations of X.500, Profile, and
DNANS should be included. The second aspect of the field
trials is to distinguish issues inherent in the X.500
specification from artifacts of a particular implementation of
it. Therefore, if possible, two implementations of X.500
should be included. Only one such implementation, Quipu, was
identified as developed enough to be included in a field trial
at present, but others are under way, and will follow. This
stage must also include a careful and objective review of the
field trials.
- Stage 2: Implementation.
This stage will include work on both the service and user
interfaces. The field trials could result in one of a variety
of conclusions about the service. These may range from
concluding that one or another of the services suits the needs
of the NRN to proposing a compromise position based on a
combination of shortcomings of any one service and the features
of others to address those shortcomings. Because X.500 will
become the standard in other domains, an interface to X.500
will be necessary. Since all of these implementations are
still under development, in order to provide production quality
code, more implementation work will be needed.
Although some work will have been done on the user interfaces,
much more will be needed in this stage to provide a variety of
interfaces. Much emphasis should be placed on this in Stage 2.
- Stage 3: Deployment.
Deployment of the full white pages service requires information
gathering in order to fill the directory service, placement of
servers, distribution of and training for use of client code,
placement and management of services, and delegation of
authority within the service for authority over the contents.
Data collection and some delegation of authority as well as
training for users of the client code would begin during the
field trial. This stage would begin concurrently with the
other two. During the second year, detailed planning for
deployment must take place. This stage would conclude in three
years, at which time widespread deployment would have occurred.
In order to undertake this three stage program effectively, the group
identified the following major projects:
- Further implementation of code for the field trials.
In each case (e.g., Quipu, Profile, and DNANS), programs exist,
although modifications are likely to be necessary. For
example, each will need to be modified to utilize the common
file format into which the input data about users will be
gathered.
- Design, development and evaluation of user interfaces.
- Design and development of data gathering and management tools.
- Oversight and evaluation of the field trials.
Careful thought and planning must go into the field trials, to
guarantee that we learn what is needed to make an evaluation
and to plan for the white pages service. The evaluation must
also produce a document that is both a general specification
(assuming no one alternative is chosen wholesale) and profiling
information, in order for later interoperability and
conformance testing.
- Detailed planning and later management of deployment.
This includes delegation of authority over parts of the
namespace and arbitrating the shape of the namespace
(addressing the questions about who gets what sorts of names).
This is in addition to the continued and extended data
collection and management, distributing the data, placing the
code, documentation and user education.
- Standards participation is an important part of the program.
It is critical as X.500 changes during the next 4 year study
period that the United States take a strong stand on any
changes we envision. It is encumbant on us to utilize
effectively the results of the largest field trials of this
work in the international arena. The group agreed that this
could take up to one half of one person's time in a year.
- A task force or working group is necessary to provide a forum
for communication and discussion.
It is important to pursue this path now, both to architect a unified
solution before a collection of ad hoc solutions is deployed, and to
provide effective input into the X.500 standards work based on the
field trials.
2. Goals and Requirements for a White Pages Service
The requirements of a white pages service are the following:
- Functionality:
The simple form of a white pages service is straightforward;
one should be able to query the service with the name of a
person, and have returned attributes of the person such as
network mail address and phone number.
- Correctness of information:
The information in a white pages service is useless and
untrusted if it is not updated regularly. A white pages
service will not be used, if the information it provides is out
of date or incorrect. This will require a set of management
tools. Data integrity is an especially difficult challenge in
this area, in contrast with information that is syntactically
correct.
- Size:
The science and research community has been estimated at ten
million users. The number of organizations in the United
States is on the order of ten to one hundred thousand.
- Usage and query rate:
In comparison with the typical telephone book pattern of about
one lookup a week per person, users of electronic mail in the
science and research community will send more electronic mail
messages than they currently make phone calls, leading to an
estimate of ten searches a week per user for electronic as well
as paper mail and telephone information. This leads to a query
rate of 10**8 queries per week or 170 per second on average,
with much higher peak rates. The average could probably be
handled by a single server, but not the peak rates and this
would leave little room for growth. Therefore, a distributed,
multiple server solution is the only one that make sense.
- Response time:
The issue of overall query behavior must be considered
carefully. The issue arises when queries, in particular
searches, are not limited to tightly constrained sets of
entries. Since the number of queries generated will be
proportional to the number of users (and the size of the
system), the white pages design must avoid costs per query that
are related to the size of the system. The consequence,
otherwise, will be quadratic behavior in response time.
The response time of the service must also reflect the expected
usage. A phone book style query must respond in the waiting
time tolerable to a user, perhaps ten seconds maximum, or one
second desirable. If the service is incorporated as a
component of a larger service, then the needs of that service
determine the response time.
- Partitioned Authority:
The white pages service under discussion would be used by a
wide variety of organizations, ranging from small and large
companies, to network service providers, to government
agencies. Many of these would find it unacceptable to delegate
the authority over their namespaces to some other organization.
Therefore, partitioned authority including some access control,
name assignment, and information management must be possible.
- Access Control:
The access control required by the white pages falls into two
categories, read access control, and write or modify access
control. There are at least two reasons that read access
control must be available. One is that organizations may
require limiting the access to the actual entries or parts of
them. This would be comparable to organizations not being
willing to distribute their corporate phone books or personnel
records. The other reason is that some organizations do not
want to publicize or make public their organizational
structure. Write and modify access control is necessary
because both individuals and organizations may want to prevent
inadvertent or malicious creation or modification of
information. Access control is an issue for both organizations
wanting to retain local control of personnel information and
individuals wanting to control access to private information
about themselves.
- Multiple Transport Protocol Support:
Within the next three years, one cannot expect all the
organizations in the USA to convert to the OSI protocols. On
the other hand, some will. It is therefore important that any
white pages service provide interfaces on top of both OSI
protocols and TCP/IP. There currently exists a partial OSI
suite know as ISODE on top of TCP. This is being distributed
widely enough that perhaps this should also be supported.
In addition to these requirements, there are a number of features
that would make a white pages service more useful. These are:
- Additional Functionality:
Descriptive naming with sophisticated searching based on
attributes would support a more flexible human interface than
simple name translation. Descriptive naming also would support
a general yellow pages style service.
The form of a yellow pages service is less certain. One
definition of a yellow pages service is a directory that stores
a number of pre-computed inversions of the directory database,
so that entries can be retrieved very efficiently using these
predetermined attributes of the data. Another definition of a
yellow pages service is one that provides a very powerful set
of search primitives, somewhat in common with a relational
query language, to support retrieval of entries that match
complex attribute conditions. In other words, one view of a
yellow pages service is that it is constructed to avoid
expensive searches, the other is that it is to facilitate
general searches.
- Accountability:
Accountability is important both for allocation and recovery of
costs. Vendors may provide commercial directory services,
therefore depending on accounting as part of their successful
commercial ventures.
- Multiple Interfaces:
There should be both human and programming interfaces to the
white pages. For example, in addition to human lookups, mail
services could effectively use a naming service allow users to
include human oriented names than the current electronic mail
addresses that are required, such as full domain names.
- Multiple Clients:
Several different clients should exist both to provide for a
variety of styles of human usage, and to support selection of
the most commonly used computer environments (e.g., UNIX, VMS,
MSDOS, OS2, MAC/OS).
3. Pre-existing Services
This section identifies other naming services that have been proposed
or implemented for naming people. Implementations of all of these
exist, although some are still only experimental.
Internet Domain Naming Service
The Internet Domain Name Service [6,1] is used today to name
host machines. It is implemented to address the query rates
and database sizes consistent with looking up hosts as part of
mail delivery. It provides a hierarchy with delegation of
authority within the hierarchy. Aliases are also available.
There is no access control, and the service is widely
distributed throughout the Internet. It supports management of
distribution, replication and caching. It is operational, and
provides a rich base of practical experience. It was
originally intended to be extensible to cover naming of people.
It runs on a variety of different operating systems and
utilizes the TCP/IP protocol suite.
The DECnet Network Architecture Naming Service (DNANS)
There is a rather well developed specification [5,3] for a
naming service that is part of the DECnet architecture, which
in turn arose from work at the DEC SRC in Palo Alto. This
architecture addresses some problems not yet covered by X.500,
such as access control, replication, and caching. It was
explicitly defined to have great scalability and management
features. It provides a global hierarchy of names, which are
mapped into properties. Therefore, operations of searching
based on properties or attributes may be expensive and
difficult. At present it is only implemented on VMS using the
DNA protocols, but will be moved to UNIX and TCP in the next
year.
Clearinghouse
This service [7,2] is part of the Xerox network environment.
It operates today as a global service for Xerox. They have
considerable experience with its operation, including problems
of scale. Clearinghouse provides a three-level hierarchy of
names that are mapped to sets of properties. Loose consistency
is provided through slow propagation of updates. Both this
service and the DEC service mentioned above are to some extent
based on an earlier Xerox service called Grapevine.
Profile
A project at the University of Arizona run by Larry Peterson
[8] has produced a white pages name service called Profile. It
supports descriptive naming and sophisticated lookup tools.
Profile assumes the existence of some other service such as the
DNS to navigate among Profile servers. This navigation service
need not restrict the relationship among Profile servers to a
hierarchical organization; Profile supports a non-hierarchical
global structure. Names in Profile consist of sets of
attributes. Experimental implementations are in operation
today, and the largest site currently contains about 10,000
entries. The Profile code has been available for long enough
that it has become stable. The implementation is UNIX-based
only and uses TCP.
X.500
X.500 is the CCITT recommendation (also ISO/IEC/DIS 9594) [4]
for a directory service. Because it is a CCITT recommendation,
it evolves in four year study periods, one of which has
recently come to a close. Thus, X.500 has a stable definition
for the next four years.
In X.500, the set of all objects forms a single hierarchy, with
each object being named relative to its parent and a single
root as the topmost parent. An object consists of a set of
attributes. Searching can be done by use of a logical
combination of attribute values, known as a filter. A subset
of these attributes comprise an object's distinguished name
relative to its parent. The hierarchy as described in the
CCITT recommendation is geographic at its top level and
organizational within that. Alternatives can also be defined,
although they are not part of the CCITT or ISO documents. In
addition, there is no proposed mechanisms for distributing
information about other attribute types or object classes. As
with the other services, X.500 is a distributed service. It
specifies cooperating servers or Directory Server Agents (DSAs)
under local control and management each of which knows about
one or more parts of the hierarchy. The clients are known as
Directory User Agents (DUAs). It is defined to run on top of
the OSI protocol stack. The demonstrations of X.500 in the
context of Internet run on top of the ISODE package, which
provides OSI transport on top of TCP.
X.500 is incomplete in that there are a number of identifiable
areas in which the standard says nothing, but that need to be
specified for a successful implementation. Some examples of
these are: access control (although authentication is
supported), replication, caching, the database itself (the
shape of the hierarchy), tools to limit the scope and cost of
searching, and database management tools.
There are currently a small number of implementations of X.500
in progress at such locations as University College London (the
Quipu project, on UNIX using ISODE), the University of British
Columbia (UNIX based using their own full OSI suite), MIT
(experimental, Symbolics Lisp Machine based, Lisp using TCP),
The Wollongong Group (offshoot of Quipu), The Retix
Corporation, NIST, and at least several underway in Italy and
Japan. There are probably others and a number of other
American corporations have discussed building their own. Each
of these must make its own decision in the areas in which X.500
is silent. Quipu is probably the most complete implementation
of X.500 to date. The pilot version has about 20 DUAs in seven
countries with an estimated 20,000 entries total.
4. Proposed Approach
The conclusion of this report is that some form of X.500 is the most
likely candidate. The reasons for this decision are that it has a
rich semantics and will become the international de facto standard.
There are, however, serious problems with its incompleteness and with
its strict hierarchy. Therefore, in order to explore these and
become convinced of its viability, the attendees at the meeting
agreed on field trials, as a first stage. Initially, this would
include experiments with at least one X.500 implementation (Quipu),
Profile to explore a non-hierarchical structure and richer
descriptive naming, and DNANS in order to explore some of the
incomplete aspects of X.500 for which DNANS has architected
solutions.
A three-stage plan, with all three stages beginning coincidentally
and as soon as possible, would provide such a service within the NRN.
The first stage should be complete in a year, the second in two, and
the third in three. Stage 1 would be field trials of three
approaches to naming with an emphasis on distinguishing between the
specification and a particular implementation of X.500, as well.
Stage 2 would be a more complete implementation of a white pages
service base on the conclusions from Stage 1. Stage 3 would be
widespread deployment of the implementation developed in Stage 2.
The planning for Stage 3 is not outlined here in detail, because that
plan would be part of the proposed work to be done. If the field
trials were to lead to the conclusion that none of the services is
adequate, the plan for the remainder of the work would need to be
rescheduled.
If the Internet community is to adopt X.500 (or any other standard),
it is necessary to make a number of design and management decisions,
above and beyond the implementation decisions for the DSA. Since
there are a number of such decisions to be resolved, and some of
these are significant, the group recommended that this planning and
management function should be recognized as a distinct activity.
4.1. Stage 1: The Field Test
It was agreed that field trials would be a valuable form in which to
explore the issues of building a white pages service for two reasons.
First, the software is still in early stages of development or
deployment. Some of it is production code, but still first release;
the rest is part of research projects. Second, it is important to
learn from experience with a limited and sympathetic community. The
suggested community was the computer science community, in
particular, computer science departments. That will not be the case
completely, since the computer science community in general does not
use DECnet. Therefore, for experiments with the DNANS, the NASA/DOE
community was recommended. They will be using DNANS in any case, as
they move to DECnet Phase V.
The twofold purpose of the field trials is to explore differing
directory service architectures and to refine the study of X.500
specifically, to distinguish architectural aspects of it from
features of a particular implementation of X.500. Initially, the
trials would include the Quipu implementation of X.500, Profile, and
the DNANS. A second implementation of X.500 should be identified and
included as soon as possible. Part of the emphasis of the field
trials would be on gathering and maintenance of naming information.
To ease this, a single common file format for storage of and access
to the naming information and use of a single set of data management
tools was recommended, although no particular set was identified.
The various directory services would need to be retrofitted to this
file format. Such consistency in file format would mean that the
services could all be co-resident, sharing files, thus permitting
single locations to participate in several parts of the field trials.
This, in turn, would allow for direct comparisons.
There are a number of issues, which are not addressed in X.500, that
would need to be resolved for a large scale deployment such as a
white pages for the NRN. In particular, these are: clients of the
service; data collection and maintenance; distribution, replication
and caching of information; access control, accountability, and
information integrity; and support by non-OSI protocols. Each of the
name services included in the field trials would include decisions in
these areas, albeit different ones. The field trials will allow for
evaluation of these different mechanisms.
There are two other major issues that must also be addressed:
functionality and size. Functionality encompasses both the first
point of the nature of the interfaces to the service as well as the
structure of the namespace (e.g., hierarchy). A discussion of size
must include not only the number of entries handled by the service as
a whole, but how those entries are distributed and the query and
update patterns.
In general, all of these issues are tightly coupled, but are
separated here for the purposes of understanding the field trials and
its potential effectiveness. They would also be the issues that
would be the basis for the work done in Stage 2 of the project.
- Functionality:
X.500 and DNANS make strong statements about the organization
of the namespace. In both cases, it is a single, absolute
hierarchy with soft links or aliases and attribute-based naming
useful both in searches of subtrees of the hierarchy and for
storing information about the objects in the hierarchy. The
searches are based on logical combinations of attribute values.
Quipu implements the naming structure and search functionality
as specified in X.500. In contrast, Profile, provides a more
general facility that supports any form of relative names, not
just hierarchical, and a small programming language to express
the functions for searching. By including Profile in the field
trials, these more general facilities can be tested.
X.500 specifies that the service is separated into two parts
for implementation of the service, known as the Directory
Service Agent (DSA), and the client, known as the Directory
User Agent (DUA). DUAs can be implemented independently of the
implementation of the white pages service. Quipu, Profile, and
DNANS have taken different approaches to the presentation model
for DUAs, so the three implementations will allow for
additional experience.
- Size:
As discussed earlier, a white pages service must be prepared to
handle a minimum of 10**7 entries, although they may be
distributed, and a query rate of hundreds per second. It must
also be prepared to handle much higher peak rates. If the
address lookup that is presently provided by the DNS is also
supported by the white pages service, the query rate will be
much higher. The designers of the field trials must determine
whether or not such usage will be part of the final service and
therefore must be examined in the field trials. If so, caching
may be part of the solution. In addition, the response time
for DUAs must be reasonable for a human sitting at a console.
Furthermore, modifications to the data should occur in
reasonably short periods of time, although this could be
measured in hours.
The field trials must allow for experimentation under such
stressful conditions. The environment for testing must have
both large and small nodes, as well as both heavy and light
load querying and situations in which reorganization can be
tested. Such reorganization may be a simple as moving one
piece of the hierarchy to another point and handling naming
conflicts in the new environment. X.500 does not address this
issue, but it will be needed by the NRN.
- Distribution, replication, and caching:
These are areas in which X.500 has very little to say, but a
great deal of work has been done in other distributed, network
naming services, in particular both the DNS and DNANS. There
seems to be general agreement that distribution of naming
services should be done on the basis of nodes in the naming
structure, which also provide the basis for administrative
partitioning. All the naming services described here support
distribution, partitioning of the information for placement on
cooperating servers. Neither X.500 (and therefore Quipu) nor
Profile is prepared to redistribute portions of the namespace,
for reallocation of administrative responsibilities or load
balancing, although this should be possible and DNANS is
prepared to do so. Replication is necessary for accessibility
in a large-scale or global namespace, although again X.500 does
not address this issue. Quipu has taken a stand on this, by
defining master and slave copies of the data; it is similar to,
but not the same as, the approach taken in the DNS. Caching is
barely touched on in X.500 and not at all in Profile, but our
experience with the DNS indicates that caching is critical to
effective operation of a distributed name service. The DNANS
has an architected solution based on objects in the namespace
as the unit of distribution and replication. Again, the DNANS
solution should be explored in the field test environment.
- Access control, accountability, and integrity:
Access control and accountability require some degree of
authentication. X.500 supports authentication based on using
an RSA public key algorithm, but does not address issues of
universal registration, nor issues of access control or
accountability themselves. These are left as a local issue,
although depending on the design of the system, they may have
global implications. The problem of integrity of the
information in the name service is nowhere addressed. Profile
also does not address these issues, although it uses
authentication based on UNIX authentication, involving user ids
and passwords. DNANS takes a strong stand on access control,
architecting it in at the level of individual entries. Field
trials will force these issues out into the open.
- Structure of the naming tree:
In the deployment of the DNS, about one year was lost to
arguments about the actual structure of the naming hierarchy.
People form strong opinions about their name, and fight for or
against certain hierarchical structures. The same issue will
arise here, and advanced planning to deal with the problem is
required.
In this case, the problem is made harder by the fact that the
hierarchy will be global; X.500 is an international standard,
based on the assumption that there is only one example of the
tree, partitioned by country. Probably the American White
Pages Service, at least at its root, will be run by the NIST or
its contractor. We must deal with the problem that in the
short term, various implementations may not interwork, and we
must work with NIST to support the needed services.
Specific issues that come up related to the naming tree are:
* How is delegation of control of the tree managed?
For example, who decides what DSA holds what parts
of the tree?
* How is the creation of new parts of the tree
(e.g., an organizational entry) controlled?
- Support for Tree Search:
Regardless of the defintion of the white pages service in the
NRN, it will need to interface to the X.500 world. The X.500
naming hierarchy can be expected to become very large, and
guidance is needed for users to help them navigate the tree.
Users need help to find their way to unknown parts of the
namespace. As in other naming services, a feature of X.500 is
that additional entries, aliases (similar to links in file
systems) can be installed to provide an easy path for a user in
one part of the tree to find other interesting parts of the
tree. By establishing a consistent policy for the use of alias
entries, learning how to navigate the tree can be made much
easier for a user. As part of setting up the tree, therefore,
these sorts of policies need to be defined.
- Definition of database structures:
There are a number of data structures that need to be defined
as part of setting up each of the services. These include, for
example, the types of information stored for the entry about a
person. This information must be stored in the servers, and
passed to the clients. These structures must thus be
specified. In other words, the schema defining attributes and
object classes must be specified for the NRN.
- Load balancing:
The dynamic performance of the Internet system must be
estimated, so that the servers can be sized properly.
Especially at the root of the tree, the query rate must be
estimated carefully. Caching will have a strong influence on
this. Therefore, traffic patterns are very dependent on the
details of implementation.
- Supporting multiple protocol suites:
At least three protocol suites are and will continue to be used
in the NRN environment. They are DECnet, TCP/IP, and the OSI
suite of protocols. Since the white pages service is at the
applications layer, it must run on top of at least these three
protocol suites. It is important to understand the
requirements of the white pages service for its transport
protocols.
By addressing these issues within the field trials, we will be
preparing for the further development of Stage 2. A result of Stage
1 will be a detailed specification of the white pages service,
possibly an extension to or modification of X.500. This should
dovetail with the activities specifying the details required for
implementation (known as "profiling") by the NIST Workshop for
Implementors of OSI. In addition, in order to run the field trial,
the information capture problem must be addressed, providing the some
of the preliminary work of Stage 3.
4.2. Stage 2: Implementation
If the evaluation of Stage 1 concludes that some form of X.500 is
acceptable, at least one of the two X.500 implementations included in
the field trials should provide the basis for a production quality
implementation of X.500 for general deployment. Further work will
likely be needed on the basis of the evaluations of the field trials.
A production version of an implementation requires both reliable
servers as well as a variety of clients to provide differing
interfaces on a mixture of hardware and operating systems.
In addition, especially because of the inclusion of Profile and
DNANS, a variety of different DUAs will be explored by definition.
Further investigation into the DUAs should begin in parallel with or
in conjunction with the field trials. There should be distinct DUAs
for both programs and humans. In addition, there probably should be
human-user DUAs geared both to the naive user with simple usage
patterns and the more sophisticated user who wants to perform complex
queries. It is also important to design DUAs that do not require a
great deal of computing power for the small machines still in use in
great quantity. Much of the user community may not be able to afford
expensive equipment upgrades.
Assuming that X.500 is deemed to be the specification of the service,
the field trials will address many issues not included in X.500 as of
1989. Since it is important for the NRN to support interconnectivity
beyond its own bounds, it behooves us to feed what has been learned
back into the standards activities. This was identified as a
separate activity because of the intellectual as well as time
commitment that must be made to do this effectively.
4.3. Stage 3: Deployment
A plan is required to develop the schedule of service introduction,
and to co-ordinate the deployment as it is undertaken. This includes
mediating service problems, a significant task in its own right.
The details of deployment were not discussed at the meeting, although
several of the seeds of deployment lie in Stages 1 and 2. The first
of these is the capture and management of information. The second is
DUA development. Both of these must be included Stage 1 in order to
support a usable environment for the trials. In addition, the
information that will have been captured in Stage 1 could be printed
producing a hard copy of the white pages information. That could be
distributed to all scientists and engineers involved; such a project
would provide an early white pages service. During the initial
periods of both Stages 1 and 2, planning for deployment would also
have to proceed, in order to provide a smooth transition to this
third stage in the project.
5. Conclusion
The consensus of the meeting was that following a path that included
X.500 was both the correct direction and feasible, although X.500
needs further elaboration. There were several important items for
further study. The first is that there are many issues left
unresolved in X.500 that have been addressed in other naming
services, and the NRN should take advantage of the solutions in those
other services. The second is that there was some reservation about
certain features of X.500, especially in the area of the imposition
of a hierarchy for naming, and only limited flexibility in
descriptive naming. The participants believe that is important
understand whether X.500 provides enough mechanisms to work around
such problems by finding a higher common ground that includes the
best features of all the naming services included in the field
trials. The final issue with respect to X.500 was that there was
agreement that X.500 will be an accepted and utilized standard in at
least part of the networked community and therefore interfacing to it
will be necessary. Given that, and the other reasons for choosing
X.500, the consensus was that the plan described above would bring
the NRN and its community of users a useful and usable white pages
service.
References
1. Austein, R., "The Internet Domain Name System", Proceedings of
USA Decus, Massachusetts Institute Technology/LCS, 1987.
2. Demers, A., D. Greene, C. Hauser, W. Irish, J. Larson, S.
Shenker, H. Sturgis, D. Swinehart, and D. Terry, "Epidemic
algorithms for replicated database maintenance", Proceedings of
the 6th Symposium on Principles of Distributed Computing, ACM,
Vancouver, B.C., Canada, pp. 12-21, August 1987.
3. Digital Equipment Corporation, "DNA Naming Service Functional
Specification Version 1.0.1", Order number: EK-DNANS-FS-001,
Digital Equipment Corporation, 1988.
4. International Organization for Standardization, "Information
Processing Systems - Open Systems Interconnection - The
Directory", Draft Standard (In 8 parts), Also CCITT
Recommendation X.500, 1988.
5. Lampson, B., "Desiging a Global Name Service," Proceedings of the
5th Symposium on Principles of Distribute Computing, ACM,
Calgary, Alberta, Canada, pp. 1-10, August 1986.
6. Mockapetris, P., "Domain Names - Concept and Facilities", RFC
1034, USC/Information Sciences Institute, November 1987.
7. Oppen, D., and Y. Dalal, "The Clearinghouse: A Decentralized
Agent for Locating Named Objects in a Distributed Environment",
Tech. Rept. OPD-T8103, Xerox Corporation, Palo Alto, CA, October
1981.
8. Peterson, L., "Profile: A System for Naming Internet Resources",
Tech. Rept. 20, Department of Computer Science, University of
Arizona, June 1987.
Author's Address
Karen R. Sollins
Massachusetts Institute of Technology
Laboratory for Computer Science
545 Technology Square
Cambridge, MA 02139-1986
Phone: (617) 253-6006
EMail: SOLLINS@XX.LCS.MIT.EDU