Network Working Group D. Green
Request for Comments: 1679 P. Irey
Category: Informational D. Marlow
K. O'Donoghue
NSWC-DD
August 1994
HPN Working Group Input to the IPng Requirements Solicitation
Status of this Memo
This memo provides information for the Internet community. This memo
does not specify an Internet standard of any kind. Distribution of
this memo is unlimited.
Abstract
This document was submitted to the IETF IPng area in response to RFC
1550. Publication of this document does not imply acceptance by the
IPng area of any ideas expressed within. Comments should be
submitted to the big-internet@munnari.oz.au mailing list.
Executive Summary
The Navy's High Performance Network (HPN) working group has studied
the requirements of mission critical applications on Navy platforms.
Based on this study, three basic categories of issues for IPng have
been identified. The assumptions identified include accommodation of
current functionality, commercial viability, and transitioning. The
general requirements identified include addressing, integrated
services architecture, mobility, multicast, and rapid route
reconfiguration. Finally, the additional considerations identified
include fault tolerance, policy based routing, security, and time
synchroniztion. The HPN working group is interested in participating
with the IETF in the development of standards which would apply to
mission critical systems. In particular, the HPN working group is
interested in the development of multicast functionality, an
integrated services architecture, and support for high performance
subnetworks.
The HPN working group has been established to study future network
architectures for mission critical applications aboard Navy
platforms. As a result, the HPN working group is interested in the
results of the IPng selection and development process. This document
is a product of discussions within the HPN working group.
Green, Irey, Marlow & O'Donoghue [Page 1]
RFC 1679 HPN IPng Requirements August 1994
The purpose of this document is to provide what the HPN working group
perceives as requirements for an IPng protocol set. Many of the
necessary capabilities exist in current Internet and ISO network
protocols; however, the HPN working group has identified needed
capabilities that are beyond the existing standards.
The HPN working group has identified three categories of topics for
discussion in this document. The first category is assumptions or
those topics that the HPN working group believes the IPng process
will solve satisfactorily without specific Navy input. The second
category is general requirements. These are capabilities that are
felt to be insufficiently addressed in existing network protocols and
of key importance to Navy mission critical applications. Finally, a
set of additional considerations has been identified. These are also
issues of importance to the HPN working group. However, no guidance
or specific requests can be provided at this time.
The US Navy has set up a program through the Space and Naval Warfare
Systems Command called the Next Generation Computer Resources (NGCR)
Program. The purpose of this program is to identify the evolving
needs for information system technology in Navy mission critical
systems. The NGCR High Performance Network (HPN) working group was
recently established by the NGCR program to examine high performance
networks for use on future Navy platforms (aircraft, surface ships,
submarines, and certain shore-based applications). This working group
is currently reviewing Navy needs. The requirements provided below
are based on the HPN working group's current understanding of these
Navy application areas. The application areas of interest are further
examined below. The time frame for design, development, and
deployment of HPN based systems and subsystems is 1996 into the
twenty first century.
Three general problem domains have been identified by the HPN working
group. These are the particular problem domains within a mission
critical environment that the HPN working group is targeting. The
first is a distributed combat system environment. This problem
domain is analogous to a collection of workstations involved in many
varied applications involving multiple sources and types of
information. Analog, audio, digital, discrete, graphic, textual,
video, and voice information must be coordinated in order to present
a single concise view to a commander, operator, or any end user. The
second problem area highlights the general internetworking
environment. The task of moving information to many heterogeneous
systems over various subnetworks is addressed. Finally, the problem
of providing a high speed interconnect for devices such as sensors
and signal processors is identified. [1]
Green, Irey, Marlow & O'Donoghue [Page 2]
RFC 1679 HPN IPng Requirements August 1994
The application area of HPN is the communication network which is a
component of the mission critical systems of Navy platforms. The
expected end points or users of the HPN include humans, computers,
and the many devices (cameras, etc.) found on such platforms. The
function of these end points includes sensor input, signal
processors, operator consoles, navigation systems, etc. The endpoints
are typically grouped into systems both on platforms and at shore-
based sites. These systems perform functions including long range
planning, analysis of sensor information, and machinery control in
real-time.
Information types that have been identified as required by the HPN
working group include voice, live and pre-recorded audio ranging from
voice to CD quality (e.g., from sensors), video (1 to 30 frames per
second in both monochrome and color), image data (static or from
real-time sensors), reliable and connectionless data transfer, and
very high-bandwidth (gigabits per second) unprocessed sensor data.
Another way of categorizing the HPN application area is by
considering the user services that need to be supported. Some of
these services are the following:
1. process to process message passing
2. distributed file and database manipulation
3. e-mail (both within the platform and off the platform)
4. teleconferencing (with the platform, between platforms, and
across the Internet)
5. video monitoring of various physical environments
6. voice distribution (as a minimum between computer processes
and people)
7. image services
8. time synchronization
9. name or directory services
10. network and system management
Green, Irey, Marlow & O'Donoghue [Page 3]
RFC 1679 HPN IPng Requirements August 1994
11. security services (support of multilevel data security,
privacy and protection)
The assumptions documented below are concerns that the HPN working
group presumes will be accommodated in the IPng process. However,
they are of enough importance to this working group to merit
identification.
The IPng protocols need to provide for at least the existing
functionality. In particular, the following issues have been
identified.
1) The IPng protocols need to provide for the basic
connectionless transfer of information from one end-point to
another.
2) The IPng protocols need to support multiple subnetwork
technologies. This includes but is not limited to Ethernet,
FDDI, Asynchronous Transfer Mode (ATM), Fiber Channel, and
Scalable Coherent Interface (SCI). These are the subnetwork
technologies that are of particular interest to the HPN
working group. Ideally, IPng protocols should be subnetwork
independent.
3) The IPng protocols need to support hosts that may be
multihomed. Multihomed in this context implies that a single
host may support multiple different subnetwork technologies.
Multihomed hosts must have the capability to steer the traffic
to selected subnetworks.
4) The IPng process needs to recognize that IPng may be only one
of several network protocols that a host utilizes.
5) The IPng process needs to provide for appropriate network
management in the finished product. Network management is of
vital importance to the applications of interest to the HPN
working group.
As is the case in the commercial world, the HPN working group feels
strongly that the IPng protocols must be commercially viable. This
includes but is not limited to the following issues:
Green, Irey, Marlow & O'Donoghue [Page 4]
RFC 1679 HPN IPng Requirements August 1994
1) The IPng protocols must function correctly. The Navy cannot
afford to have network protocol problems in mission critical
systems. There must be a high degree of confidence that the
protocols are technically sound and multi-vendor
interoperability is achievable.
2) The IPng protocols must have the support of the
commercial/industrial community. This may first be
demonstrated by a strong consensus within the IETF community.
The Navy has a large number of existing networks including both
Internet and ISO protocols as well as a number of proprietary
systems. As a minimum, the IPng effort must address how to
transition from existing IP based networks. Additionally, it would be
desirable to have some guidance for transitioning from other network
protocols including, but not limited to, CLNP and other commonly used
network protocols. The transition plan for IPng needs to recognize
the large existing infrastructure and the lack of funds for a full
scale immediate transition. There will, in all likelihood, be a long
period of co-existence that should be addressed.
The general requirements documented below are topics that the HPN
working group considers to be of vital importance in a network
protocol solution. It is hoped that the IPng solution will address
all of these issues.
The HPN working group has identified initial addressing requirements.
First, a large number of addresses are required. In particular, the
number of addressable entities on a single platform will range from
the 100's to 100,000. The number of large platforms (ships,
submarines, shore based sites) will range from a few hundred to
several thousand. In addition, there will be 500 to 1000 or more
small platforms, primarily aircraft. Since it is expected that in
the future many of these platforms will be connected to global
networks, the addresses must be globally unique.
The second requirement identified is for some form of addressing
structure. It is felt that this structure should be flexible enough
to allow for logical structures (not necessarily geographical) to be
applied. It is also felt that this is important for the
implementation of efficient routing solutions. In addition, the
addressing structure must support multicast group addressing. At a
Green, Irey, Marlow & O'Donoghue [Page 5]
RFC 1679 HPN IPng Requirements August 1994
minimum 2**16 globally unique multicast groups must be
distinguishable per platform.
An important goal of the HPN working group is to identify existing
and emerging technologies which provide mechanisms for integrating
the services required by mission critical Navy systems. The HPN
working group has identified two classes of problems under the
general category of integrated services. The first is to provide for
the multiple types of services identified in section 2.1. It is
required to support these services in an integrated fashion in order
to be able to correlate (in time) related streams of information.
The second class of problems relates to the predictable management of
the various traffic flows associated with the above identified
services. While many of these services require the delivery of a PDU
within a specified time window, the applications in a mission
critical environment can demand more stringent requirements. In areas
where real-time systems are in use, such as machinery control,
narrower and/or more predictable delivery windows may be required
than in the case of the delivery of audio or video streams. The
mission critical environment also requires the ability to assign
end-to-end importance to instances of communications (i.e.,
invocations of a particular service). For example, an ongoing video
stream may need to yield to machinery control commands to ensure that
the commands are received before their deadline. The expense of this
action is to degrade temporarily the video stream quality.
The HPN working group is looking for mechanisms in the IPng protocols
to provide for both of these classes of problems in an integrated
fashion. An integrated services architecture reduces design and
integration complexities by providing a uniform set of tools for use
by the mission critical system designer and application developer.
Finally, the integrated services architecture must be flexible and
scalable so that new services can be added in the future with minimum
impact on systems using it. The HPN working group has intentionally
avoided mentioning particular mechanisms that can be used to solve
some of these problems in order to avoid requiring a particular
solution.
The HPN working group has identified two classes of mobility for the
Navy mission critical environment. First, most platforms are
themselves mobile. As these platforms move from port to port or from
flight deck to flight deck, it is important that they are able to
communicate with a number of defense installations via a general
Green, Irey, Marlow & O'Donoghue [Page 6]
RFC 1679 HPN IPng Requirements August 1994
infrastructure. Additionally, it is feasible that systems within a
single platform may be mobile. Maintenance and damage assessment
requires large amounts of information at numerous locations on a
platform. This information could possibly be made available through
mobile terminals.
Multicast transfer is a very critical IPng requirement for the Navy's
mission critical systems. Aboard a Naval platform there are many
hosts (e.g., workstations) connected via numerous subnetworks. These
hosts are all working different aspects of the problem of keeping the
platform operational to perform its mission. In support of this
environment, multicast transfer is needed to share data that is
needed by multiple hosts. For example, aboard a ship platform,
environmental data (roll, pitch, heading...) is needed by almost all
systems. Video conferencing may be used for communication among
operational personnel at multiple places aboard this ship. Video
conferencing could also be used for communicating with personnel on
other platforms or at shore facilities. Both of these examples, in
addition to a number of DoD and NATO studies, have highlighted the
need for multicast functionality in mission critical systems.
One of the limiting factors with the present IP version 4 multicast
is the optional nature of this multicast, particularly with respect
to routers. The use of tunnels, while enabling the initial deployment
of multicast in the Internet, appears to limit its potential. The HPN
working group believes that the best approach to provision of
multicast functionality is to consider it as a basic functionality to
be provided by IPng. In addition, sensible mechanisms are needed to
control multicast traffic (i.e., scope control). Finally, support is
required to enable multicast functionality in IPng in areas such as
group addressing and scalable multicast routing.
The HPN project will be using very high bandwidth subnetwork
technology. In the mission critical environment one very important
problem is placing a very low bound on the time it takes to identify
a subnetwork problem and to complete the necessary route
reconfigurations. The Navy's mission critical environment needs to be
able to trade-off bandwidth to enable a short
detection/reconfiguration time on subnetwork faults. A maximum bound
on this time is felt to be less than 1 second.
Green, Irey, Marlow & O'Donoghue [Page 7]
RFC 1679 HPN IPng Requirements August 1994
This section represents additional concerns of the mission critical
environment which may impact IPng. The HPN working group felt that
these issues are important for the mission critical environment;
however, it was not clear how or whether it is necessary to
accommodate them in IPng solutions. It may suffice that designers of
IPng are aware of these issues and therefore do not preclude
reasonable solutions to these problems.
The mission critical environment is particularly sensitive to the
area of fault tolerance. Any mechanisms that can be accommodated
within the IPng protocol set, including routing and management, to
support various levels of fault tolerance are desirable. In
particular, the following features should be supported: error
detection, error reporting, traffic analysis, and status reporting.
The HPN working group feels that there may be some uses for policy
based routing within the Navy's mission critical systems. The
primary interest is in support of a very capable security facility.
Other uses discussed are as a means for keeping certain types of data
on certain subnetworks (for multiply homed hosts) and providing for
automatic reconfiguration in the event of particular subnetwork
failures.
Security is an important requirement for most Navy applications and
thus the ability for the network functions to be designed to support
security services are essential. The following are several security
services in particular that the HPN working group believes the
network function should be able to support: rule based access
control, labeling, authentication, audit, connection oriented and
connectionless confidentiality, selective routing, traffic flow
confidentiality, connection oriented and connectionless integrity,
denial of service protection, continuity of operations, and
precedence/preemption. In addition to these services, the network
function should also support the security management of these
security services. In particular, key management is of importance.
Currently, the IPSEC of the IETF has several draft memos being
considered to incorporate various security services in the network
functions. It is of concern to the HPN working group that the IPng be
able to support the concepts currently being developed by the IPSEC
Green, Irey, Marlow & O'Donoghue [Page 8]
RFC 1679 HPN IPng Requirements August 1994
and also provide the ability for the addition of future security
services.
Time synchronization among the various components of mission critical
systems is of vital importance to the Navy. It is desirable to be
able to synchronize systems on multiple subnetworks via a network
layer infrastructure. Some hooks for time synchronization can be
envisioned in the network layer. However, the HPN working group
feels that, as a minimum, efficient time synchronization algorithms
must be able to function above an IPng infrastructure. For HPN
systems, it is desirable that a time-of-day synchronization
capability be supported of at least an accuracy of one microsecond
among all hosts in a platform or campus network. The IPng protocols
should not arbitrarily prevent this type of synchronization
capability.
A number of concerns specific to mission critical systems targeted by
the HPN working group have been identified. The HPN working group is
interested in participating with the IETF in the development of
standards which would apply to mission critical systems. In
particular, the HPN working group is interested in the development of
multicast functionality, an integrated services architecture, and
support for high performance subnetworks.