∂21-JAN-75  1337		network site RAND
 To:  F. Corbato, J. Markowitz, J. McCarthy, M. Pirtle,
         T. Stockham, I. Sutherland, B. Woods
 
 From:  Peter Weiner  (WEINER @ RAND-RCC)
 
 Subject: Issues for the Working Group on Computing
                 Resource Procurement
 
 ------
 
 
 (I have asked Ivan Sutherland to join the Working Group.
         His NET address will be SUTHERLAND @ RAND-RCC.)
 
 
 As I noted in my invitation to you, ARPA has not provided
 our group with a tightly defined charter;  rather we must
 define this for ourselves through the selection of the set
 of questions/issues that we will address.
 
 I have prepared the following in order to stimulate discussion.
 I am not sure that we will want to cover all of the following,
 and there must be issues of importance not listed.
 
 In preparing the list, I have assumed that we will look
 at both large and small machine procurement, and that we will
 look at the full spectrum of ARPA/IPTO activities (e.g.,
 6.1 basic research, 6.2 advanced development, technology transfer
 programs, etc.).  Thus some items may have to to be examined in
 several contexts.
 
 While I am sure that some of you will want to add items to the
 list, it is important that we don't try to cover too much ground.
 (Perhaps this list is already far too extensive - what do you think?)
 Consequently, please offer your thoughts regarding the relative
 importance of what follows, so it may be appropriately pruned.
 
 
 1.  Near-term expansion of facilities
 
         What machines should be used to add capability to
         today's computing environment within the next year
         or so?  Specifically, what do we think of the KL-10.
         Are there developments on the horizon (say the
         IBM Future System) that suggest ARPA should defer
         procurement at this time.
 
 
 2.  General Configuration
 
         How will networks influence the general computing
         environment in 5-10 years.  What will the role
         be of general time-shared machines?  machines
         dedicated to classes of computations?  front-end
         machines  and intelligent terminals?
 
 
 3.  Hardware and Software Standards
 
         Should new systems be acquired from a limited number of
         manufacturers or should a wide diversity of capability
         be encouraged?
 
         Should new systems be software compatible?  Is
         standardization of operating systems important or
         feasible?
 
 
 4.  Inter-machine Communication
 
         Which machines should communicate through a
         common network?  through dial-up phone lines?
         For what purposes?  mail only?  remote computation?
         distributed computing?
 
         Should communication be provided to machines or data bases
         not directly used for ARPA research?  which? for what
         purposes?
 
         Is distributed computing (i.e. where one asks the network
         for service, not a particular machine) technically
         and economically feasible as a means to satisfy
         a significant portion of the computing needs of the
         ARPA research community.  When can this be realized?
 
 
 5.  Hardware Development and Maintenance
 
         Will there be a need within the research community
         for ARPA to sponsor the design and construction
         of hardware such as computers specialized to handle
         problems of importance to the community (e.g. large
         lisp programs)?
 
         Are mechanisms needed to insure that ARPA developed
         hardware is commercially available and maintainable?
         What mechanisms?
 
 
 6.  Software Development and Maintenance
 
         Should ARPA sponsor the development of supporting
         software?  What kind?  Operating systems?  Languages
         and compilers/interpreters?  Utilities?
 
         Who should maintain ARPA sponsored software?
 
 
 7.  Location and Operational Management of ARPA Machines
 
         Should all machines be physically located at one
         site?  Several sites (how many)?  Where?
 
         Who runs the machines?  Service organizations that do
         not conduct research?  Non-university research
         organizations?  University research groups?
 
         What services should be provided?
 
 
 8.  Financial
 
         How should ARPA finance computer acquisition?  Buy?
         Lease? Rent?
 
         Should the purchase of computer service from commercial
         companies be limited to incremental short-term needs,
         or an important component of ARPA's computing procurement.
 
         How much flexibility should individual contractors
         have in selecting the supplier of computing service?
 
         What accounting mechanisms should be used to charge
         research contractors for computer resource utilization?
 
 
 
 
 9.  National Software Works
 
         Should the NSW or NSW-like protocols be used to tie
         together the software and hardware systems of ARPA
         research contractors?
 
 
 10. Secure Computing
 
         What provisions should be made to allow classified
         computing and the storage of classifed documents?
         What provisions should be made to allow storage and
         transmission of sensitive non-classified documents.
 %

∂21-JAN-75  1239		1,PAW
 to fly from kyoto to pisa you would fly from tokyo to rome, there are no
 direct flights, JAL has a flight that makes one stop...the flying time is
 approximately 15 hours and the cost (if the ticket were purchased here) is
 $942.00.....Patte