As per Relevance of the word research, we have this rfc below:
Network Working Group Gigabit Working
Request for Comments: 1077 B. Leiner,
November 1988
Critical Issues in High Bandwidth
Status of this
This memo presents the results of a working group on High
Networking. This RFC is for your information and you are
to comment on the issues presented. Distribution of this memo
unlimited
At the request of Maj. Mark Pullen and Maj. Brian Boesch of DARPA,
ad-hoc working group was assembled to develop a set
recommendations on the research required to achieve a
high-bandwidth network as discussed in the FCCSET recommendations
Phase III
This report outlines a set of research topics aimed at providing
technology base for an interconnected set of networks that
provide highbandwidth capabilities. The suggested research
draws upon ongoing research and augments it with basic and
components. The major activities are the development
demonstration of a gigabit backbone network, the development
demonstration of an interconnected set of networks with
throughput and appropriate management techniques, and the
and demonstration of the required overall architecture that
users to gain access to such high bandwidth
Gigabit Working Group [Page 1]
RFC 1077 November 1988
1. Introduction and
1.1.
The computer communications world is evolving toward both high
bandwidth capability and high-bandwidth requirements. The
workshop conducted under the auspices of the FCCSET Committee on
Performance Computing [1] identified a number of areas
extremely high-bandwidth networking is required to support
scientific research community. These areas range from
graphical visualization of supercomputer results through the
of high rate sensor data from space to the ground-based
investigator. Similar requirements exist for other applications
such as military command and control (C2) where there is a need
quickly access and act on data obtained from real-time sensors.
workshop identified requirements for switched high-bandwidth
in excess of 300 Mbit/s to a single user, and the need to
service in the range of a Mbit/s on a low-duty-cycle basis
millions of researchers. When added to the needs of the military
commercial users, the aggregate requirement for
service adds up to many billions of bits per second. The results
this workshop were incorporated into a report by the FCCSET [2].
Fortunately, technology is also moving rapidly. Even today,
installed base of fiber optics communications allows us to
aggregate bandwidths in the range of Gbit/s and beyond to
geographical regions. Estimates arrived at in the workshop lead
to believe that there will be available raw bandwidth
terabits per second
The critical question to be addressed is how this raw bandwidth
be used to satisfy the requirements identified in the workshop: 1)
provide bandwidth on the order of several Gbit/s to individual users
and 2) provide modest bandwidth on the order of several Mbit/s to
large number of users in a cost-effective manner through
aggregation of their traffic
Through its research funding, the Defense Advanced Research
Agency (DARPA) has played a central role in the development
packet-oriented communications, which has been of tremendous
to the U.S. military in terms of survivability and interoperability
DARPA-funded research has resulted in the ARPANET, the first packet
switched network; the SATNET, MATNET and Wideband Network,
demonstrated the efficient utilization of shared-access
channels for communications between geographically diverse sites
Gigabit Working Group [Page 2]
RFC 1077 November 1988
packet radio networks for mobile tactical environments; the
and TCP/IP protocols for interconnection and interoperability
heterogeneous networks and computer systems; the development
electronic mail; and many advances in the areas of network security
privacy, authentication and access control for distributed
environments. Recognizing DARPA's past accomplishments and
desire to continue to take a leading role in addressing these issues
this document provides a recommendation for research topics
gigabit networking. It is meant to be an organized compendium of
critical research issues to be addressed in developing the
base needed for such a high bandwidth ubiquitous network
1.2. Ongoing
The OSTP report referred to above recommended a three-phase
to achieving the required high-bandwidth networking for
scientific and research community. Some of this work is now
underway. An ad-hoc committee, the Federal Research
Coordinating Committee (FRICC) is coordinating the interconnection
the current wide area networking systems in the government;
those of DARPA, Department of Energy (DoE), National
Foundation (NSF), National Aeronautics and Space
(NASA), and the Department of Health and Human Services (HHS).
accordance with Phases I and II of the OSTP report, this
will provide for an interconnected set of networks to
research and other scholarly pursuits, and provide a basis for
networking for this community. The networking is being
through shared increased bandwidth (current plans are to share a 45
Mbit/s backbone) and coordinated interconnection with the rest of
world. In particular, the FRICC is working with the
networking community under the auspices of another ad-hoc group,
Coordinating Committee for Intercontinental Research
(CCIRN), to establish effective US-Europe networking
However, as the OSTP recommendations note, the required bandwidth
the future is well beyond currently planned public, private,
government networks. Achieving the required gigabit
capabilities will require a strong research activity. There
considerable ongoing research in relevant areas that can be
upon; particularly in the areas of high-bandwidth
links, high-speed computer switching, and high-bandwidth local
networks. Appendix A provides some pointers to current
efforts
Gigabit Working Group [Page 3]
RFC 1077 November 1988
1.3. Document
This report outlines a set of research topics aimed at providing
technology base for an interconnected set of networks that
provide the required high-bandwidth capabilities discussed above
The suggested research focus draws upon ongoing research and
it with basic and applied components. The major activities are
development and demonstration of a Gigabit Backbone network (GB) [3],
the development and demonstration of an interconnected set
networks with gigabit throughput and appropriate
techniques, and the development and demonstration of the
overall architecture that allows users to gain access to such
bandwidth. Section 2 discusses functional and performance
along with the anticipated benefits to the ultimate users of such
system. Section 3 provides the discussion of the critical
issues needed to achieve these goals. It is organized into the
areas of technology that need to be addressed: general
issues, high-bandwidth switching, high-bandwidth host interfaces
network management algorithms, and network services. The
in some cases contains examples of ongoing relevant research
potential approaches. These examples are intended to clarify
issues and not to propose that particular approach. A discussion
the relationship of the suggested research to other
activities and optimal methods for pursuing this research is
in Section 4.
2. Functional and Performance
In this section, we provide an assessment of the types of services
GN (four or five orders of magnitude faster than the
networks) should provide to its users. In instances where we
there would be a significant impact on performance, we have
an estimate of the amount of bandwidth needed and delay allowable
provide these services
2.1. Networking Application
It is envisioned that the GN will be capable of supporting all of
following types of networking applications
Gigabit Working Group [Page 4]
RFC 1077 November 1988
Currently Provided Packet
It is important that the network provide the users with
equivalent of services that are already available in packet
switched networks, such as interactive data exchange,
service, file transfer, on-line access to remote
resources, etc., and allow them to expand to other more
services to meet their needs as they become available
Multi-Media
This capability will allow users to take advantage of
media types (e.g., graphics, images, voice, and video as well
text and computer data) in the transfer of messages,
increasing the effectiveness of message exchange
Multi-Media
Such conferencing requires the exchange of large amounts
information in short periods of time. Hence the requirement
high bandwidth at low delay. We estimate that the bandwidth
range from 1.5 to 100 Mbit/s, with an end-to-end delay of no
than a few hundred msec
Computer-Generated Real-time
Visualizing computer results in the modern world of
requires large amounts of real time graphics. This in turn
require about 1.5 Mbit/s of bandwidth and no more than
hundred msec. delay
High-Speed Transaction
One of the most important reasons for having an ultra-high-
network is to take advantage of supercomputing capability.
are several scenarios in which this capability could be utilized
For example, there could be instances where a non-
may require a supercomputer to perform some processing and
some intermediate results that will be used to perform
further processing, or the exchange may be between
supercomputers operating in tandem and periodically
results, such as in a battle management, war gaming, or
control applications. In such cases, extremely short
times are necessary to accomplish as many as hundreds
interactions in real time. This requires very high bandwidth,
the order of 100 Mbit/s, and minimum delay, on the order
hundreds of msec
Gigabit Working Group [Page 5]
RFC 1077 November 1988
Wide-Area Distributed Data/Knowledge Base Management
Computer-stored data, information, and knowledge is
around the country for a variety of reasons. The ability
perform complex queries, updates, and report generation as
many large databases are one system would be extremely powerful
yet requires low-delay, high-bandwidth communication
interactive use. The Corporation for National
Initiatives (NRI) has promoted the notion of a National
base with these characteristics. In particular, an
approach is to cache views at the user sites, or close by to
efficient repeated queries and multi-relation processing
relations on different nodes. However, with caching, a
activity may incur a miss in the midst of a query or update
causing it to be delayed by the time required to retrieve
missing relation or portion of relation. To minimize the
for cache directories, both at the server and client sites,
unit of caching should be large---say a megabyte or more.
addition, to maintain consistency at the caching client sites
server sites need to multicast invalidations and/or updates
Communication requirements are further increased by replication
the data. The critical parameter is latency for cache misses
consistency operations. Taking the distance between sites to
on average 1/4 the diameter of the country, a one Gbit/s data
is required to reduce the transmission time to be roughly the
as the propagation delay, namely around 8 milliseconds for
size of unit. Note that this application is supporting far
sophisticated queries and updates than normally associated
transaction processing, thus requiring larger amount of data to
transferred
2.2. Types of Traffic and Communications
Different types of traffic may impose different constraints in
of throughput, delay, delay dispersion, reliability and
delivery. Table 1 summarizes some of the main characteristics
several different types of traffic
Gigabit Working Group [Page 6]
RFC 1077 November 1988
Table 1: Communication Traffic
+------------------------+-------------+-------------+-------------+
| | | | Error-free |
| Traffic | Delay | Throughput | Sequenced |
| Type | Requirement | Requirement | Delivery |
+------------------------+-------------+-------------+-------------+
| Interactive Simulation | Low |Moderate-High| No |
+------------------------+-------------+-------------+-------------+
| Network Monitoring | Moderate | Low | No |
+------------------------+-------------+-------------+-------------+
| Virtual Terminal | Low | Low | Yes |
+------------------------+-------------+-------------+-------------+
| Bulk Transfer | High | High | Yes |
+------------------------+-------------+-------------+-------------+
| Message | Moderate | Moderate | Yes |
+------------------------+-------------+-------------+-------------+
| Voice |Low, constant| Moderate | No |
+------------------------+-------------+-------------+-------------+
| Video |Low, constant| High | No |
+------------------------+-------------+-------------+-------------+
| Facsimile | Moderate | High | No |
+------------------------+-------------+-------------+-------------+
| Image Transfer | Variable | High | No |
+------------------------+-------------+-------------+-------------+
| Distributed Computing | Low | Variable | Yes |
+------------------------+-------------+-------------+-------------+
| Network Control | Moderate | Low | Yes |
+------------------------+-------------+-------------+-------------+
The topology among users can be of three types: point-to-point (one
to-one connectivity), multicast (one sender and multiple receivers),
and conferencing (multiple senders and multiple receivers).
are three types of transfers that can take place among users.
are connection-oriented network service, connectionless
service, and stream or synchronous traffic. Connection
connectionless services are asynchronous. A connection-
service assumes and provides for relationships among the
packets sent over the connection (e.g., to a common destination
while connectionless service assumes each packet is a complete
separate entity unto itself. For stream or synchronous service
reservation scheme is used to set up and guarantee a constant
steady amount of bandwidth between any two subscribers
Gigabit Working Group [Page 7]
RFC 1077 November 1988
2.3. Network
The GB needs to be of high bandwidth to support a large population
users, and additionally to provide high-speed connectivity
certain subscribers who may need such capability (e.g., between
supercomputers). These users may access the GN from local
networks (LANs) directly connected to the backbone or via high-
intermediate regional networks. The backbone must also
end-to-end delay to support highly interactive high-
(supercomputer) activities
It is important that the LANs that will be connected to the GN
permitted data rates independent of the data rates of the GB.
speeds should be allowed to change without affecting the GB, and
GB speeds should be allowed to change without affecting the LANs.
this way, development of the technology for LANs and the GB
proceed independently
Access rate requirements to the GB and the GN will vary depending
user requirements and local environments. The users may
access rates ranging from multi-kbit/s in the case of terminals
personal computers connected by modems up to multi-Mbit/s and
for powerful workstations up to the Gbit/s range for high-
computing and data resources
2.4. Directory
Directory services similar to those found in CCITT X.500/ISO DIS 9594
need to be provided. These include mapping user names to
mail addresses, distribution lists, support for
checking, access control, and public key encryption schemes
multimedia mail capabilities, and the ability to keep track of
users (those who move from place to place and host computer to
computer). The directory services may also list facilities
to users via the network. Some examples are databases
supercomputing or other special-purpose applications, and on-
help or telephone hotlines
The services provided by X.500 may require some extension for GN
For example, there is no provision for multilevel security, and
approach taken to authentication must be studied to ensure that
meets the requirements of GN and its user community
Gigabit Working Group [Page 8]
RFC 1077 November 1988
2.5. Network Management and
The objective of network management is to ensure that the
functions smoothly and efficiently, and consists of the following
accounting, security, performance monitoring, fault isolation
configuration control
Accounting ensures that users are properly billed for the
that the network provides. Accounting enforces a tariff; a
expresses a usage policy. The network need only keep track of
items addressed by the tariff, such as allocated bandwidth, number
packets sent, number of ports used, etc. Another type of
may need to be supported by the network to support resource sharing
namely accounting analogous to telephone "900" numbers.
accounting performed by the network on behalf of resource
and consumers is a pragmatic solution to the problem of getting
users and consumers into a financial relationship with each
which has stymied previous attempts to achieve widespread use
specialized resources
Performance monitoring is needed so that the managers can tell
the network is performing and take the necessary actions to keep
performance at a level that will provide users with
service. Fault isolation using technical control mechanisms
needed for network maintenance. Configuration management allows
network to function efficiently
Several new types of routing will be required by GN. In addition
true type-of-service, needed to support diverse
applications, real-time applications, interactive applications,
bulk data transfer, there will be need for traffic controls
enforce various routing policies. For example, policy may
that traffic from certain users, applications, or hosts may not
permitted to traverse certain segments of the network
Alternatively, traffic controls may be used to promote fairness;
is, to make sure that busy link or network segment isn't dominated
a particular source or destination. The ability of applications
reserve network bandwidth in advance of its use, and the use
strategies such as soft connections, will also require development
new routing algorithms
2.6. Network Security
Security is a critical factor within the GN and one of those
that are difficult to provide. It is envisioned that
Gigabit Working Group [Page 9]
RFC 1077 November 1988
unclassified and classified traffic will utilize the GN,
protection mechanisms must be an integral part of the network
strategy. Features such as authentication, integrity
confidentiality, access control, and nonrepudiation are essential
provide trusted and secure communication services for network users
A subscriber must have assurance that the person or system he
exchanging information with is indeed who he says he is
Authentication provides this assurance by verifying that the
source of a query request, control command, response, etc., is
actual source. Integrity assures that the subscriber's
(such as requests, commands, data, responses, etc.) is not changed
intentionally or unintentionally, while in transit or by replays
earlier traffic. Unauthorized users (e.g., intruders or
viruses) would be denied use of GN assets through access
mechanisms which verify that the authenticated source is
to receive the requested information or to initiate the
command. In addition, nonrepudiation services can be offered
assure a third party that the transmitted information has not
altered. And finally, confidentiality will ensure that the
of a message are not divulged to unauthorized individuals
Subscribers can decide, based upon their own security needs
particular activities, which of these services are necessary at
given time
3. Critical Research
In the section above, we discussed the goals of a research program
gigabit networking; namely to provide the technology base for
network that will allow gigabit service to be provided in
effective way. In this section, we discuss those issues which
feel are critical to address in a research program to achieve
goals
3.1. General Architectural
In the last generation of networks, it was assumed that bandwidth
the scarce resource and the design of the switch was dictated by
need to manage and allocate the bandwidth effectively. The
basic change in the next generation network is that the speeds of
trunks are rising faster than the speeds of the switching elements
This change in the balance of speeds has manifested itself in
ways. In most current designs for local area networks,
Gigabit Working Group [Page 10]
RFC 1077 November 1988
bandwidth is not expensive, the design decision was to trade
effective use of the bandwidth for a simplified switching technique
In particular, networks such as Ethernet use broadcast as the
distribution method, which essentially eliminates the need for
switching element
As we look at still higher speed networks, and in particular
in which the bandwidth is still the expensive component, we
design new options for switching which will permit effective use
bandwidth without the switch itself becoming the bottleneck
The central thrust of new research must thus be to explore
network architectures that are consistent with these very
speed assumptions
The development of computer communications has been
distorted by the characteristics of wide-area networking:
high cost, low speed, high error rate, large delay. The time is
for a revolution in thinking, technology, and approaches,
to the revolution caused by VCR technology over 8 and 16 mm.
technology
Fiber optics is clearly the enabling technology for high-
transmission, in fact, so much so that there is an expectation
the switching elements will now hold down the data rates.
conventional circuit switching and packet switching have
problems at higher data rates. For instance, circuit
requires increasing delays for FTDM synchronization to handle skew
In the case of packet switching, traditional approaches require
much processing per packet to handle the tremendous data flow.
problem for both switching regimes is the "intelligence" in
switches, which in turn requires electronics technology
Besides intelligence, another problem for wide-area networks
storage, both because it ties us to electronics (for the
future) and because it produces instabilities in a large-
system. (See, for instance, the work by Van Jacobson on self
organizing phenomena for self-destruction in the Internet.)
Techniques are required to eliminate dependence on storage, such
cut-through routing
Overall, high-speed WANs are the greatest agents of change,
greatest catalyst both commercially and militarily, and the area
for revolution. Judging by the attributes of current high-
network research prototypes, WANs of the future will be photonic
multi-gigabit networks with enormous throughput, low delay, and
error rate
Gigabit Working Group [Page 11]
RFC 1077 November 1988
A zero-based budgeting approach is required to develop the new high
speed internetwork architecture. That is, the time is ripe
significantly rethink the Internet, building on experience with
system. Issues of concern are manageability,
evolvability and support for the new communication requirements
including remote procedure call, real-time, security and fault
tolerance
The GN must be able to deal with two sources of high-
requirements. There will be some end devices (computers)
more or less directly to the GN because of their
requirements for high bandwidth (e.g., supercomputers needing
drive remote high-bandwidth graphics devices). In addition,
aggregate traffic due to large numbers of moderate rate
(estimates are roughly up to a million potential users needing up
1 Mbit/s at any given time) results in a high-bandwidth
in total on the GN. The statistics of such traffic are different
there are different possible technical approaches for dealing
them. Thus, an architectural approach for dealing with both must
developed
Overall, the next-generation architecture has to be, first
foremost, a management architecture. The directions in link speeds
processor speeds and memory solve the performance problems for
communication situations so well that manageability becomes
predominant concern. (In fact, fast communication makes
systems more prone to performance, reliability, and
problems.) In many ways, the management system of the
is the ultimate distributed system. The solution to this
problem may well require the best talents from the communications
operating systems and distributed systems communities, perhaps
drawing on database and parallelism research
3.1.1. High-Speed Internet using High-Speed
The GN will need to take advantage of a multitude of different
heterogeneous networks, all of high speed. In addition to
based on the technology of the GB, there will be high-speed LANs.
key issue in the development of the GN will be the development of
strategy for interconnecting such networks to provide gigabit
on an end to end basis. This will involve techniques for switching
interfacing, and management (as discussed in the sections below
coupled with an architecture that allows the GN to take
advantage of the performance of the various high-speed networks
Gigabit Working Group [Page 12]
RFC 1077 November 1988
3.1.2. Network
The GN will need an architecture that supports the need to manage
system as well as obtain high performance. We note that almost
human-engineered systems are hierarchically structured from
standpoint of control, monitoring, and information flow.
hierarchical design may be the key to manageability in the next
generation architecture
One approach is to use a general three-level structure,
to interadministrational, intraadministrational, and
networks. The first level interconnects communication facilities
truly separate administrations where there is significant
of security, accounting, and goals. The second level
subadministrations which exist for management convenience in
organizations. For example, a research group within a university
function as a subadministration. The cluster level consists
networks configured to provides maximal performance among hosts
are in frequent communication, such as a set of diskless
and their common file server. These hosts are typically, but
necessarily, geographically collocated. For example, two
networks may be tightly coupled by a fiber optic link that
between the two physical networks, making them function as one
Research along these lines should study the
characteristics of communications, such as those being
by the IAB Task Force on Autonomous Networks. Based on
results, we expect that such work would clearly demonstrate
considerable communication takes place between
subadministrations in different administrations;
patterns are not strictly hierarchical. For example, there might
intense direct communication between the experimental
departments of two independent universities, or between the
support group of one company and the operating system
group of another. In addition, (sub)administrations may well
require divisions into public information and private information
3.1.3. Fault-Tolerant
Although the GN will be developed as part of an experimental
program, it will also serve as part of the infrastructure
researchers who are experimenting with applications which will
such a network. The GN must have reasonably high availability
support these research activities. In addition to facilitate
transfer of this technology to future operational military
Gigabit Working Group [Page 13]
RFC 1077 November 1988
commercial users, it will need to be designed to become
reliable. This can be accomplished through diversity of
paths, the development of fault-tolerant switches, use of
distributed control structure with self-correcting algorithms,
the protection of network control traffic. The architecture of a
should support and allow for all of these things
3.1.4. Functional Division of Control Between Network
Current protocol architectures use the layered model of
decomposition first developed in the early work on ARPANET protocols
The concept of layering has been a powerful concept which has
dramatic variation in network technologies without requiring
complete reimplementation of applications. The concept of
has had a first-order impact on the development of
standards for data communication---witness the ISO "Reference
for Open Systems Interconnection."
Unfortunately, however, the powerful concept of layering has
paired, both in the DoD Internet work and the ISO work, with
extremely weak concept of the interface between layers.
interface designs are all organized around the idea of commands
responses plus an error indicator. For example, the TCP
interface provides the user with commands to set up or close a
connection and commands to send and receive datagrams. The user
well "know" whether they are using a file transfer service or
character-at-a- time virtual terminal, but can't tell the TCP.
underlying network may "know" that failures have reduced the path
the user's destination to a single 9.6 kbit/s link, but it also can'
tell the TCP implementation
All of the information that an analyst would consider crucial
diagnosing system performance is carefully hidden from
layers. One "solution" often discussed (but rarely implemented)
to condense all of this information into a few bits of "Type
Service" or "Quality of Service" request flowing in one
only---from application to network. It seems likely that
approach cannot succeed, both because it applies too much
to the knowledge available and because it does not provide two-
flow
We believe it to be likely that the next-generation network
require a much richer interface between every pair of adjacent
if adequate performance is to be achieved. Research is needed
the conceptual mechanisms, both indicators and controls, that can
implemented at these interfaces and that, when used, will result
Gigabit Working Group [Page 14]
RFC 1077 November 1988
better performance. If real differences in performance can
observed, then the implementors of every layer will have a
incentive to make use of the mechanisms
We can observe the first glimmers of this sort of
between layers in current work. For example, in the ISO work
are 5 classes of transport protocol which are supposed to provide
range of possible matches between application needs and
capabilities. Unfortunately, it is the case today that the class
transport protocol is chosen statically, by the implementer,
than dynamically. The DARPA Wideband net offers a choice of
or datagram service, but typically a given host uses all one or
the other---again, a static rather than a dynamic choice.
research that we believe is needed, therefore, is not how to
alternatives, but how to provide them and choose among them on
dynamic, real-time basis
3.1.5. Different Switch
One approach to high-performance networking is to design a
that is expected to work as a stand-alone demonstration,
addressing the need for interconnection to other networks. Such
experiment may be very valuable for rapid exploration of the
space. However, our experience with the Internet project
that a primary research goal should be the development of a
architecture that permits the interconnection of a number
different switching technologies
The Internet project was successful to a large extent because
could incorporate a number of new and preexisting
technologies: various local area networks, store and
switching networks, broadcast satellite nets, packet radio networks
and so on. In this way, it decoupled the use of the protocols from
particular technology base. In fact, the technology base
rapidly, but the Internet protocols themselves provided a
that led to their success
The next-generation architecture must similarly deal with a
and evolving technology base. We see "fast-packet" switching
being developed (for example in B-ISDN); we see photonic
and wavelength division multiplexing as more advanced technologies
We must divorce our architecture from dependence on any one of these
At the host interface, we must divorce the multiplexing of the
from the form of data that the host sees. Today the packet is
both as multiplexing and interface element. In the future, the
Gigabit Working Group [Page 15]
RFC 1077 November 1988
may see the network as a message-passing system, or as memory.
the same time, the network may use classic packets,
division, or space division switching
A number of basic functions must be rethought to provide
architecture that is not dependent on the underlying switching model
For example, our transport protocols assume that data will be lost
units of a packet. If part of a packet is lost, we discard the
thing. And if several packets are systematically lost in sequence
we may not recover effectively. There must be a host-level unit
error recovery that is independent of the network. This sort
abstraction must be applied to all the aspects of
specification: error recovery, flow control, addressing, and so on
3.1.6. Network Operations, Monitoring, and
There is a hierarchy of progressively more effective
sophisticated techniques for network management that
regardless of network bandwidth and application considerations
1. Reactive problem
2. Reactive resource
3. Proactive problem
4. Proactive resource management
Today's network management strategies are primarily reactive
than proactive: Problem management is initiated in response to
complaints about service outages; resource allocation decisions
made when users complain about deterioration of quality of service
Today's network management systems are stuck at step 1 or
step 2 of the hierarchy
Future network management systems will provide proactive
management---problem diagnosis and restoral of service before
become aware that there was a problem; and proactive
management---dynamic allocation of network bandwidth and
resources to ensure that an acceptable level of service
continuously maintained
The GN management system should be expected to provide
problem and resource management capabilities. It will have to do
while contending with three important changes in the managed
environment
Gigabit Working Group [Page 16]
RFC 1077 November 1988
1. More complicated devices under
2. More diverse types of
3. More variety of application protocols
Performance under these conditions will require that we
re-think how a network management system handles the expected
volumes of raw management-related data. It will become
important for the system to provide thresholding, filtering,
alerting mechanisms that can save the human operator from drowning
data, while still permitting access to details when diagnostic
fault isolation modes are invoked
The presence of expert assistant capabilities for early
detection, diagnosis, and problem resolution will be mandatory
These capabilities are highly desirable today, but they will
essential to contend with the complexity and diversity of devices
applications in the Gigabit Network
In addition to its role in dealing with complexity,
provides the only hope of controlling and reducing the high costs
daily management and operation of a GN
Proactive resource management in GNs must be better understood
practiced, initially as an effort requiring human intervention
direction. Once this is achieved, it too must become automated to
high degree in the GN
3.1.7. Naming and Addressing
Current networks, both voice (telephone) and data, use
structures which closely tie the address to the physical location
the network. That is, the address identifies a physical
point, rather than the higher-level entity (computer, process, human
attached to that access point. In future networks, this
aspect of addressing must be removed
Consider, for example, finding the desired party in the
network of today. For a person not at his listed number, finding
number of the correct telephone may require preliminary calls,
which advice is given to the person placing the call. This
well when a human is placing the call, since humans are well
to cope with arbitrary conversations. But if a computer is
the call, the process of obtaining the correct address will have
be incorporated in the architecture as a core service of the network
Gigabit Working Group [Page 17]
RFC 1077 November 1988
Since it is reasonable to expect mobile hosts, hosts that
connected to multiple networks, and replicated hosts, the issue
mapping to the physical address must be properly resolved
To permit the network to maintain the dynamic mapping to
physical address, it is necessary that high-level entities have
name (or logical address) that identifies them independently
location. The name is maintained by the network, and mapped to
current physical location as a core network service. For example
mobile hosts, hosts that are connected to multiple networks,
replicated hosts would have static names whose mapping to
addresses (many-to-one, in some cases) would change with time
Hosts are not the only entities whose physical location varies
Users' electronic mail addresses change. Within distributed systems
processes and files migrate from host to host. In a
environment where robustness and survivability are important,
applications may move about, or they may be redundant
The needed function must be considered in the context of the
and address resolution rates if all addresses in a global
network were of this sort. The distributed network
discussed elsewhere in this report should be designed to provide
necessary flexibility, and responsiveness. The nature
administration of names must also be considered
Names that are arbitrary or unwieldy would be barely better than
addresses used now. The name space should be designed so that it
easily be partitioned among the agencies that will assign names.
structure of names should facilitate, rather than hinder, the
function. For example, it would be hard to optimize the
function if names were flat and unstructured
3.2. High-Speed
The term "high-speed switching" refers to changing the switching at
high rate, rather than switching high-speed links, because the
is not difficult at low speeds. (Consider, for example,
switching of fiber connections). The switching regime chosen for
network determines various aspects of its performance, its
policies, and even its effective capabilities. As an example of
latter, it is difficult to expect a circuit-switched network
provide strong multicast support
A major area of debate lies in the choice between packet
and circuit switching. This is a key research issue for the GN
Gigabit Working Group [Page 18]
RFC 1077 November 1988
considering also the possibility of there being combinations of
two approaches that are feasible
3.2.1. Unit of Management vs.
With very high data rates, either the unit of management
switching must be larger or the speed of the processor elements
management and switching must be faster. For example, at a gigabit
a 576 byte packet takes roughly 5 microseconds to be received so
packet switch must act extremely fast to avoid being the
delay in packet times. Moreover, the storage time for the packet
a conventional store and forward implementation also becomes
significant component of the delay. Thus, for packet switching
remain attractive in this environment, it appears necessary
increase the size of packets (or switch on packet groups), do so
called virtual cut-through and use high-speed routing techniques
such as high-speed route caches and source routing
Alternatively, for circuit switching to be attractive, it
provide very fast circuit setup and tear-down to support the
nature of most computer communication. This problem is
difficult (and perhaps impossible for certain traffic loads)
the delay across the country is so large relative to the data rate
That is, even with techniques such as so-called fast select
bandwidth is reserved by the circuit along the path for almost
the propagation time before being used
With gigabit circuit switching, because it is not feasible
physically switch channels, the low-level switching is likely
FTDM on micro-packets, as is currently done in telephony.
FTDM at gigabit data rates is a challenging research problem if
skew introduced by wide-area communication is to be handled
reasonable overhead for spacing of this micro-packets. Given
lead and resources of the telephone companies, this area
investigation should, if pursued, be pursued cooperatively
3.2.2. Bandwidth Reservation
Some applications, such as real-time video, require sustained
data rate streams over a significant period of time, such as
if not hours. Intuitively, it is appealing for such applications
pre-allocate the bandwidth they require to minimize the
load on the network and guarantee that the required bandwidth
available. Research is required to determine the merits of
Gigabit Working Group [Page 19]
RFC 1077 November 1988
reservation, particular in conjunction with the different
technologies. There is some concern to raise that
reservation may require excessive intelligence in the network
reducing the performance and reliability of the network.
addition, bandwidth reservation opens a new option for denial
service by an intruder or malicious user. Thus, investigations
this area need to proceed in concert with work on
technologies and capabilities and security and
requirements
3.2.3. Multicast
It is now widely accepted that multicast should be provided as
user-level service, as described in RFC 1054 for IP, for example
However, further research is required to determine the best way
support this facility at the network layer and lower. It is
clear that the GN will be built from point-to-point fiber links
do not provide multicast/broadcast for free. At the
conservative extreme, one could provide no support and require
each host or gateway simulate multicast by sending multiple
individually addressed packets. However, there are
advantages to providing very low level multicast support (besides
obvious performance advantages). For example, multicast routing in
flooding form provides the most fault-tolerant, lowest-delay form
delivery which, if reserved for very high priority messages,
a good emergency facility for high-stress network applications
Multicast may also be useful as an approach to defeat
analysis
Another key issue arises with the distinction between so-called
group multicast and closed group multicast. In the former, any
can multicast to the group, whereas in the latter, only members
the group can multicast to it. The latter is easier to support
adequate for conferencing, for example. However, for more client
server structured applications, such as using file/database server
computation servers, etc. as groups, open multicast is required
Research is needed to address both forms of multicast. In addition
security issues arise in controlling the membership of
groups. This issue should be addressed in concert with work
secure forms of routing in general
Gigabit Working Group [Page 20]
RFC 1077 November 1988
3.2.4. Gateway
With the wide-area interconnection of local networks by the GN
gateways are expected to become a significant performance
unless significant advances are made in gateway performance.
addition, many network management concerns suggest putting
functionality (such as access control) in the gateways,
increasing their load and the need for greater capacity. This
then raise the issue of the trade-off between general-
hardware and special-purpose hardware
On the general-purpose side, it may be feasible to use a general
purpose multiprocessor based on high-end microprocessors (perhaps
exotic as the GaAs MIPS) in conjunction with a high-speed
transfer bus, as proposed as part of the FutureBus standard (which
extendible to higher speeds than currently commercially planned)
intelligent high-speed network adaptors. This would also allow
direct use of hardware, operating systems, and software
developed as part of other DARPA programs, such as
Computing. It also appears to make this gateway software
portable to commercial machines as they become available in
performance range
The specialized hardware approach is based on the assumption
general-purpose hardware, particularly the interconnection bus
cannot be fast enough to support the level of performance required
The expected emphasis is on various interconnection
techniques. These approaches appear to require greater expense,
commercial availability and more specialized software. They need
be critically evaluated with respect to the general-purpose
hardware approach, especially if the latter is using multiple
for fault-tolerance as well as capacity extension (in the absence
failure).
The same general-purpose vs. special-purpose contention is an
with operating system software. Conventionally, gateways
specialized run-time executives that are designed specifically
the gateway and gateway functions. However, the
sophistication of the gateway makes this approach less feasible.
appears important to investigate the feasibility of using a
operating system foundation on the gateways that is known to
the required security and reliability properties (as well as real
time performance properties).
Gigabit Working Group [Page 21]
RFC 1077 November 1988
3.2.5. VLSI and Optronics
It appears fairly clear that gigabit communication will use
optics for at least the near future. Without major advances
optronics to allow effectively for optical computers,
must cross the optical-electronic boundary two or more times.
are significant cost, performance, reliability, and security
for minimizing the number of such crossings. (As an example of
security benefit, optics is not prone to electronic surveillance
jamming while electronics clearly is, so replacing an optic
electronic-optic node with a pure optic node eliminates
vulnerability point.)
The benefits of improved technology in optronics is so great that
application here is purely another motivation for an already
research area (that deserves strong continued support). Therefore
we focus here in the issue of matching current (and near-
expected) optronics capabilities with network requirements
The first and perhaps greatest area of opportunity is to
totally (or largely) photonic switches in the network
nodes. That is, most packets would be switched without crossing
optics-electronics boundary at all. For this to be feasible,
switch must use very simple switching logic, require very
storage and operate on packets of a significant size. The source
routed packet switches with loopback on blockage of
illustrate the type of techniques that appear required to
this goal
Research is required to investigate the feasibility of
implementation of switches. It appears highly likely that
will at some point in the future be totally photonically switched
having the impact on networking comparable to the effect
integrated circuits on processors and memories
A next level of focus is to achieve optical switching in the
case in gateways. One model is a multiprocessor with an
interconnect. Packets associated with established paths through
gateway are optically switched and processed through
interconnect. Other packets are routed to the multiprocessor
crossing into the electronics domain. Research is required to
the networking requirements and technology with optronics technology
pushing the state of the art in both areas in the process
Given the long-term presence of the optic-electronic boundary
improvements in technology in this area are also important. However
it appears that there is already enormous commercial
Gigabit Working Group [Page 22]
RFC 1077 November 1988
activity in this area, particularly within the telephone companies
This is another area in which collaborative investigation appears
better than an new independent research effort
VLSI technology is an established technology with active
support. The GN effort does not appear to require major
initiatives in the VLSI area, yet one should be open to
novel opportunities not identified here
3.2.6. High-Speed Transfer
To achieve the desired speeds, it will be necessary to rethink
form of protocols
1. The simple idea of a stateless gateway must be replaced by
more complex model in which the gateway understands
desired function of the end point and applies
optimizations to the flow
2. If multiplexing is done in the time domain, the elements
multiplexing are probably so small that no
processing can be performed on each individually. They
be processed as an aggregate. This implies that the unit
multiplexing is not the same as the unit of processing
3. The interfaces between the structural layers of
communication system must change from a
command/response style to a richer system which
indications and controls
4. An approach must be developed that couples the
management in the host and the structure of the
data, to allow efficient transfers into host memory
The result of rethinking these problems will be a new style
communications and protocols, in which there is a much higher
of shared responsibility among the components (hosts, switches
gateways). This may have little resemblance to previous work
in the DARPA or commercial communities
3.3. High-Speed Host
As networks get faster, the most significant bottleneck will turn
to be the packet processing overhead in the host. While this
Gigabit Working Group [Page 23]
RFC 1077 November 1988
not restrict the aggregate rates we can achieve over trunks,
prevents delivery of high data rate flows to the host-
applications, which will prevent the development of new
needing high bandwidth. The host bottleneck is thus a
impediment to networked use of supercomputers
To build a GN we need to create new ways for hosts and their
bandwidth peripherals to connect to networks. We believe
pursuing research in the ways to most effectively isolate host
LAN development paths from the GN is the most productive way
proceed. By decoupling the development paths, neither is
by the momentary performance of capability bottlenecks of the other
The best context in which to view this separation is with the
of a network front end (NFE). The NFE can take the electronic
data at many data rates and transform it into gigabit light
appropriately packetized to traverse the GN. The NFE can
inputs from many types of gateways, hosts, host peripherals, and
and provide arbitration and path set-up facilities as needed.
importantly, the NFE can perform protocol arbitration to
upward compatibility with the existing Internet protocols
enabling those sophisticated network input sources to execute
specific high-throughput protocols. Of course, this introduces
need for research into high-speed NFEs to avoid the NFE becoming
bottleneck
3.3.1. VLSI and Optronics
In a host interface, unless the host is optical (an unlikely
in the near-term), the opportunities for optronic support
limited. In fact, with a serial-to-parallel conversion on
stepping the clock rate down by a factor of 32 (assuming a 32-
data path on the host interface), optronic speeds are not required
the immediate future
One exception may be for encryption. Current VLSI implementations
standard encryption algorithms run in the 10 Mbit/s range.
implementation of these encryption techniques and
techniques specifically oriented to, or taking advantage of,
capabilities appears to be an area of some potential (and
benefit if achieved).
The potential of targeted VLSI research in this area appears
for similar reasons discussed above with its application in high
speed switching. The major benefits will arise from work that
well-motivated by other research (such as high-
parallelism) and by strong commer