20 years later - where are we?

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



The 20th anniversary of the first meeting of the IETF is coming up fast - January 2006. I think a little history is in order for some of the newcomers.

I was recently cleaning out some of my older files and found the attached "Internet Problem Reports" from late 1986. These were really the first formal items the IETF took on as engineering issues. Note the assumption that TCP/IP was just a passing fancy and that ISO would be the stack going forward. Little did we know...

It is nice to know that at least a few things were eventually solved by the IETF (DNS deployment, replacement of EGP as the major routing system). And bandwidth is no longer a problem (mostly).

The attached are a scanned, OCR'd and edited to clean up format version of the original reports. Enjoy.
Appendix 1. IETF Internet Problem Descriptions

At the first IETF meeting in January 1986, a list of Internet problems
was developed covering short, intermediate and long range issues. At
the most recent IETF meeting in February 1987, an attempt was made to
develop such a list in a more rigorous fashion. The IETF membership
was divided into groups with the goal of compiling problem
descriptions in particular areas. The resulting Internet Problem
Descriptions are contained in this appendix and are a mixture of
intermediate range protocol issues and very short range O&M issues.

Problems were listed in the following format: 
Problem Description:
Severity:	(low, medium, high)
Time Frame:	(time until problem becomes critical)
Owner:	(Responsible Agency or group) 
Plan/Options:
The original forms have been edited to combine or eliminate redundant descriptions. 
The problem list is not exhaustive and further work will
1) develop a more complete list,
2) divide into categories by timeframe and 
3) prioritize within category.



IETF Internet Problem Descriptions
Problem Description:

Internet doesn't work under heavy load (ie, Congestion) For example,
existing DDN protocols can't efficiently handle gateways between
networks of grossly different band- width e.g.., Ethernet- Arpanet)

Severity: High
Time Frame: Immediate
Owner: DDN/DARPA
Plan/Options:

Short term: Add capacity to existing infrastructure

Intermediate term:
1} Develop congestion control for DoD LP
2) Investigate existing solutions outside the DDN community.
Long Term Research: Look at new Internet schemes; e.g., Internet 
Connection Oriented Protocol



IETF Internet Problem Descriptions
Problem Description:

1) Lack of ISO Connection-Less Internet Protocol in current Internet
Gateways.

2) Lack of ES-IS


Severity: Low now, grows to severe in 2 years 
Time Frame: 2 years
Plan/Options:

1) Set/define "standards" for how ISO IP should be used
2a) Start funding contractors to implement ISO/IP in gateways 
2b) Purchase gateways with ISO/IP
3) Deploy in Internet Infrastructure starting in 6 - 18 months
4) Run some applications (FTAM, etc) to gain experience. Modify 
standards goto 2)
Also: Work with Standard's Organization to apply DoD IT experience 
into ISO/IP



IETF Internet Problem Descriptions


Problem Description:

MILNET domain adoption plan

Severity: Low - now; Medium - 6 mo; High - 1-2 years 
Time Frame: (see Severity)
Owner: DDN/OSD
Plan/Options: Plans needed for vendor documentation and advice, administrator 
documentation, migration plan and RFC updates.



IETF Internet Problem Descriptions
Problem Description:

Short-term Internet Routing Problems; eg, Extra-Hop, table space
(routing) performance, buffering limitations in LSI-11, mail bridges
(gateways)

Severity: High
Time Frame: Immediate Owner: 
DDN/BBN Plan/Options:
1) Deploy Butterfly Mailbridge Gateways in Parallel with LSI-11 GW's in about 6 
months
2) Transition Core to Butterflies MB's 
3) Remove LSI-11's

Requires SW/HW to be deployed before configuration mgt and testing is completed.

IETF Internet Problem Descriptions
Problem Description:

Internet Information Management; eg, Much duplication; needless
distribution info; congestion problems

Severity: 
Time Frame: 
Owner: DDN
Plan/Options:

- reconvene the group "!%@" include regional NIC reps - look at what
  info is needed
- look at what is duplicated
- create info "way stations"
- share tools; techniques
- keep general centers informed of who to hand off users to -
  distribute data collection
- SRI-NIC acts as reference and replicates data strategically

Basic model - interlibrary loan system for traditional libraries;
everybody contributes; everybody wins; nobody pays too much or foots
the whole bill; some systems are shared others are translated; the
general NICs hand off to the specialized ones

Coordinate host liaison, host administrators, etc., by holding
meetings; getting input for net administrators



IETF Internet Problem Descriptions
Problem Description:

Name Servers

(1) Get root servers off heavily loaded hosts 
(2) See that name servers are well distributed
(3) Migration of name service to login hosts (service then part of
    backbone service) and (equipment maintained by backbone) 

Severity: Medium
Time Frame: 6 months
Owner: DDN
Plan/Options:

(1) Can negotiate immediately to get servers off heavily loaded hosts;
    evenly distributed throughout net

(2) NIC can coordinate Berkeley to get good BIND (UNIX/VAX version) of
    domain service

(3) Bring NIC into BARNET so we are on NSF net

(4) Need more capacity in login hosts; needs $ but easy to solve



IETF Internet Problem Descriptions
Problem Description:

     No organization exists to attend to problems which transcend
network boundaries, Internet O&M is not defined

Severity: High
Time Frame: Immediate
Owner: NSF/DARPA/NASA/DDN
Plan/Options:

(1) Define network and Internet O&M at next IETF 
(2) Determine organization suitable to do O&M 
(3) Draft RFC defining Internet O&M


IETF Internet Problem Descriptions
Problem Description:

IOP Facility in PSN 6 can drop messages. The current IOP module in PSN
Release 6 behaves very much like a gateway; if an 1822 host sends
messages faster than a standard X.25 host can receive them, some
percentage of the messages will be dropped. The impact of this feature
on future Internet performance should be considered.

Severity: Low


Time Frame: (Next PSN Release) 
Owner: DDN/BBN Plan/Options:

a) Determine whether this feature will exist in future PSN releases.

b) If so, evaluate potential impact on Internet performance as
   standard X.25 gateways are more widely deployed.



IETF Internet Problem Descriptions
Problem Description:

Lack of protocol testing is a severe problem in gateways and
hosts. Incompatible implementations abound.

Severity: High
Time Frame: Immediate
Owner: OSD/DCA
Plan/Options:
1) Accept the situation - ISO is coming anyway 
2) Establish testing center(s) funded by

   a gov't
   b vendors
   c private enterprise

3) If none of 2) can get funded then spend money on advertising who
the apparent "winners" are anyway; i.e., let the marketplace decide

Problem Description:

Networking research must continue into the forseeable future.  Should
its operational base be TCP/IP or ISO? TCP/IP is more accessible for
manipulation, but IS will be more prevalent and thus more realistic in
terms of providing the problems to be researched.  But will ISO
implementations be "modifiable" for/by researchers?  and how will
vendors track the research?

Severity: High
Time Frame: 5 years 
Owner: IAB


Plan/Options: Establish a study group IETF to outline the problem and 
report to all interested parties: gov't, researchers, vendors, users. While 
this looks like it overlaps with FCCSET, if they don't succeed in 
addressing it, the problem won't disappear.

IETF Internet Problem Descriptions
Problem Description:

Although several agencies have cross-country trunks, some of these are
seriously congested while others are unused. Sharing of under-utilized
trunking may help solve network congestion.


Severity: ARPA 10 
          NASA 0 
	  NSF 2


Time Frame: Immediate
Owner: TAB
Plan/Options: Interagency agreements? IRI?

IETF Internet Problem Descriptions
Problem Description:

Procedures for making changes in DDN and the internet are too
cumbersome; eg,

*	Line-in/ line-out coordination;
*	line-at-a-time acquisition leasing wastes available leverage;
*	new nodes, new hosts, additional circuits.

Severity: High
Time Frame: Immediate 
Owner: DECCO
Plan/Options: Review of current administration procedures by 
sponsoring agencies. Develop new management organization. Study 
NASA trunking concept.



IETF Internet Problem Descriptions
Problem Description:

Insufficient processing and memory capacity at some some Arpanet
PSNs. Several sites are either memory or CPU-bound because of the
growth of users and gateways

Severity: High
Time Frame: Immediate Owner: DDN
Plan/Options: Upgrade approporiate nodes from C30E's to C300's. The
sites are
*	SRI 51,
*	ISI 27,
*	RCC 5,
*	WIS 94

IETF Internet Problem Descriptions
Problem Description:

Insufficient cross-country bandwidth on ARPANET.  Highly utilized
lines induce retransmissions at the store and forward level resulting
in long delays for traffic between the two coasts. This in turn
increases the congestion and resource use seen at the source and
destination of the traffic.


Severity: High
Time Frame: Immediate 
Owner: DDN
Plan/Options: Install 2 new cross-country links: 
o MIT44 - SRI151; 
o ISI22 - Columbia (APL)

IETF Internet Problem Descriptions
Problem Description:

Internet audit trail/billing sharing 

Severity: Low
Time Frame:
Owner: JAB


Plan/Options: This is probably part of larger Network Operations. Toward this end 
- We can share audit trail/billing system
- Cooperate in building a useful interagency billing system
- Make capacity planning reports available for 
  ARPANET, MILNET, etc.


IETF Internet Problem Descriptions
Problem Description:

EGP

Severity: High
Time Frame: 6-12 Months 
Owner: DDN/DARPA
Plan/Options: Draft of EGP2 by Jose Rodrigues (SDC) and Mike 
StJohns (DDN) for next IETF

IETF Internet Problem Descriptions

Problem Description:

EGP Topology Restrictions

A) Common metric

B) Core gateway computation load

C) Information hiding by cores leads to lost 
information and suboptimal routes

D) Political restrictions - autonomy 

Severity: High (very important to NSF) 
Time Frame: Immediate
Owner: NSF/DARPA

Plan/Options:

1) Remove 3rd party routing restrictions

2) Increase base of trusted gateways/ autonomous systems 
Likeliest is new (unspecified) protocol



IETF Internet Problem Descriptions
Problem Description:
 Gateway authentications 
  A) What's a real gateway?
  B) What routes can a gateway advertise?

Severity: Low
Time Frame: 2 years 
Owner: OSD, IETF


Plan/Options: Non-authenticated gateways present denial-of-service
threats, as well as wiretapping traffic


IETF Internet Problem Descriptions
Problem Description:

Interior Gateway Protocol Problems; eg, A) GGP traffic volume B) GGP/EGP 
interactions  C) Common metrics, algorithmically converted to 
EGP common metric  D) Current IGP's not published (RIP, SPF, CISCO) 

Severity: Medium
Time Frame: 12-24 Months
Owner: IETF, DDN, BBN
Plan/Options: 
  1. Document existing IGP's
  2. Define standard (suggested/example) IGP

IETF Internet Problem Descriptions
Problem Description:

Mail Bridges 1) Administrative restrictions/routing interactions 2)
Name servers use Mail Bridges

Severity: Medium 
Time Frame: 12 year
Owner: DDN
Plan/Options: When Mail Bridges are shut down to non-mail transit traffic, 
there will be a furor aimed at DCA.

IETF Internet Problem Descriptions
Problem Description:

Gateway Redirection. Intermediate gateway decides that an alternate
route is better, has no way to inform previous gateway.

Severity: Medium
Time Frame: 12-24 Months
Owner: DDN/DARPA
Plan/Options: Develop improved Internet routing/ICMP model.

_______________________________________________

Ietf@xxxxxxxx
https://www1.ietf.org/mailman/listinfo/ietf

[Index of Archives]     [IETF Annoucements]     [IETF]     [IP Storage]     [Yosemite News]     [Linux SCTP]     [Linux Newbies]     [Fedora Users]