| Domain Name Services (DNS) and
Service Location Protocol (SLP) are both name-to-address resolution
technologies. Whether you use DNS, SLP, or some other technology
depends on the specific task at hand.
In making the choice, it is helpful to separate the tasks dealing
with name-to-address resolution from those dealing with dynamic
service discovery. The following sections discuss recommendations
for implementing name-to-address resolution, ad-hoc service discovery
and assisted service discovery.
Name-to-Address Resolution
Users invoke name-to-address resolution technologies when they
know the name of a service they wish to use but do not know the
network address of that service. Name-to-address resolution is the
primary purpose of DNS; however, a number of other technologies
are also used today to determine a network address from its name.
These technologies include Novell® Directory Services® (NDS®),
host files, SLP and Dynamic Host Configuration Protocol (DHCP).
DNS has several important advantages:
- Predictable.
- Reliable.
- Scalable to global proportions.
- Works with a variety of connections, including intranet, Internet,
dial-in and VPN.
- Leverages global infrastructure and expertise.
In addition, all the basic infrastructure and services offered
by Novell products are fully functional when DNS is used as the
name-to-address resolution mechanism.
However, DNS has two inherent disadvantages:
- Requires manual configuration.
- Requires the use of full DNS names, that is, names must include
all domain subparts.
SLP on the other hand provides:
- Automatic and dynamic configuration.
- Support for short names, that is, names need not have domain
subparts.
However, SLP has a disadvantage:
- It is limited in use to small, reliable sub-nets.
As a result, Novell recommends using DNS as the primary name-to-address
resolution mechanism and SLP as a secondary name-to-address resolution
mechanism. Novell also recommends not propagating SLP data between
remote sites if it can be avoided. That's because using remote
SLP information can lower reliability and predictability in the
system, and it may generate unnecessary traffic between remote sites.
For Ad-Hoc Service Discovery
Ad-hoc service discovery permits users to find services on the
network with little or no prior configuration required. A service
advertises its presence on the network allowing users to discover
that service without having prior knowledge of it and without administrator
intervention. Examples of technologies that support ad-hoc service
discovery are SLP, Service Advertising Protocol (SAP), Jini, Salutations
and Simple Service Discovery Protocol (SSDP).
Novell recommends SLP as the primary ad-hoc service discovery mechanism
for IP networks. SLP is particularly effective when used to discover
services that are in close proximity to the requesting user.
Novell recommends that SLP-based advertising be confined to operate
within limited geographical areas, that is, you should not replicate
SLP information across remote sites. For sites with large numbers
of advertising entities, you should deploy an SLP Directory Agent
(DA) to improve the efficiency of handling SLP requests. You may
also deploy a second SLP DA at a site to provide fault tolerance.
Using SLP for ad-hoc service discovery has two advantages:
- Dynamic configuration.
- Elegant integration with non-Novell SLP implementations.
SLP does, however, have two disadvantages:
- Non-guaranteed service lists.
- Unreliability and chattiness in global implementations.
For Assisted Service Discovery
Assisted service discovery requires network administrators to perform
some kind of setup to enable users to find services. With their
understanding of how users search for certain services, network
administrators can configure the service information to facilitate
searches. This logical organization of service data by administrators
enables users to discover-reliably and easily-the services
available to them both locally and globally. Technologies used to
support assisted service discovery include NDS, DNS and Web pages.
NDS eDirectory provides powerful search capabilities. NDS trees
constructed with logical or geographical organization are particularly
well suited for searches. Network administrators can improve the
quality of searches by populating NDS objects' attributes
with meaningful information, such as the objects' "location"
attribute.
The Web is also well suited for global searches. Most users are
familiar with the different search engines available on the Web
and the processes required to advertise sites. An important advantage
of the Web is that it offers network administrators great flexibility
in configuring Web pages to guide users through customized searches.
Location-based service discovery though the Web is an example of
using the Web for assisted service discovery.
Assisted service discovery mechanisms provide three major advantages:
- Predictable.
- Reliable.
- Scalable to global proportions.
These techniques, however, all have
a drawback:
- Static (manual) configuration.
|