REPORT OF THE COMPUTER STUDY PANEL ON CONTAINMENT OF ODP COMPUTER SPACE GROWTH IN THE EXISTING AND NEW HEADQUARTERS BUILDINGS
Document Type:
Collection:
Document Number (FOIA) /ESDN (CREST):
CIA-RDP87M00539R000901160019-2
Release Decision:
RIPPUB
Original Classification:
C
Document Page Count:
29
Document Creation Date:
December 23, 2016
Document Release Date:
January 2, 2013
Sequence Number:
19
Case Number:
Publication Date:
February 15, 1985
Content Type:
REPORT
File:
Attachment | Size |
---|---|
![]() | 1.11 MB |
Body:
ii ?i ii i ~
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
15 FEBRUARY 1985
3
CENTRAL INTELLIGENCE AGENCY
REPORT OF THE
COMPUTER STUDY PANEL
ON
CONTAINMENT OF ODP COMPUTER
SPACE GROWTH IN THE EXISTING
.AND NEW HEADQUARTERS BUILDINGS
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
STAT
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
REPORT OF THE
CENTRAL INTELLIGENCE AGENCY
COMPUTER STUDY PANEL
ON
CONTAINMENT OF ODP COMPUTER
SPACE GROWTH IN THE EXISTING
AND NEW HEADQUARTERS BUILDINGS
15 FEBRUARY 1985
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
CENTRAL INTELLIGENCE AGENCY
COMPUTER STUDY PANEL
STAT
The above named Panel members participated in the preparation of this
Report and have concurred in its contents.
STAT
III
UNCLASSIFIED
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
TABLE OF CONTENTS
PAGE
EXECUTIVE SUMMARY ................................................................................ ES-1
1. INTRODUCTION .......................................................................................... 1
2. BACKGROUND ............................................................................................ 2
3. TECHNICAL OPTIONS .............................................................................. 3
4. MANAGEMENT OPTIONS ........................................................................ 8
5. NEW BUILDING OPTIONS ........................................................................ 9
6. FINDINGS AND RECOMMENDATIONS .............................................. 19
UNCLASSIFIED
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
EXECUTIVE SUMMARY
Introduction
A Panel, convened at the direction of the CIA Deputy Director for Administration,
studied the planned expansion of Office of Data Processing (ODP) computer systems
with the goal of determining the most feasible and cost-effective method to prevent
the continuing displacement of personnel in the Headquarters building.
The Baseline information which provided the foundation for the study included:
? ODP's estimates of ADP system growth through the year 2001.
? ODP and Office of Communications (OC) plans for satisfying that growth.
? Plans for allocation and use of computer-grade space in the new Headquarters
building addition.
None of the basic assumptions in the ODP plans were challenged by the Panel;
therefore, the findings and recommendations in this report are based on the premise
that ODP's growth projections are reasonable and that their plans for satisfying the
demand are similarly reasonable ... or, at least, that alternatives which might be
better on some dimensions did not differ materially with respect to space needs.
Findings
? ODP plans to retain 36,000 sq. ft. of computer space in the present
Headquarters, even after the computer-grade space in the new addition
becomes available. The Panel sees no benefit in this approach, and urges that
all ODP computer space in the present Headquarters be released. This would
assure effective use of available space in the new addition, which has
TEMPEST features, and is specifically designed for machines.
? The computer-grade space available to ODP (150,000 sq. ft.) in the new
Headquarters addition could satisfy all growth requirements up to the year
1995, but not without displacing people. Initially, (1987), ODP plans to use
60,000 sq. ft. of this space for computers, with the remainder used for ODP of-
fice space. As computer space needs grow, the ODP people would be displaced
to make room for the computers.
? Improved technology (such as disk drives with greater storage density and
more powerful computers) make better use of space. More improvements are
forecast. ODP plans include the acquisition of this improved equipment as it
becomes available. These measures can slow the rate at which computers will
displace personnel, but not stop it.
? As long as the Agency was committed to locating all its computers in the
Headquarters building, there was no alternative but displacement of people to
accommmodate computer growth. The displaced people could be readily
accommodated elsewhere, as is evidenced by the 30-x- off-campus locations for
offices.
ES-1
UNCLASSIFIED
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
? While there may have been valid reasons in the past for not doing so, today it is
technically possible to relocate all or part of ODP's computer files and
processing capability to a remote site-preferably in the nearby Maryland or
Virginia areas.
? The security, communications, and operational problems associated with such
a remote operation are manageable, although the risks would be high if the en-
tire ODP processing capability were to be located to a remote off-campus site.
? Dividing ODP services into a local and a remote operation will have a major
impact on the ADP system architecture, but not to the extent that technical
risks are high. There is a significant peripheral benefit from a divided
operation. With proper planning and design, the remote facility and the
Headquarters facility can back each other up, providing protection against a
catastrophe at either site.
Recommendations
The Panel recommends that:
? The Agency set a ceiling of 100,000 sq. ft. of space for ODP computers in the
Headquarters compound, which would be obtained from the computer-grade
space in the new Headquarters building addition;
? The 47,462 sq. ft. of computer space in the present Headquarters building be
released for use as office space;
? An expandable, computer-grade facility of initially 50,000 sq. ft. plus support
space be provided (purchased or leased) off-site by 1992 for ODP computers
and peripheral equipment;
? The off-site facility be located at any one of several possible sites in nearby
Maryland or Virginia;
? The off-site facility be used to provide services which are less communications-
intensive (i.e., Community, development, and possibly batch applications);
? The off-site facility be linked to the Headquarters building via redundant,
secure, high bandwidth communications;
? Both the off-site facility and the Headquarters building contain computer
systems sufficiently similar to each other to provide mutual backup;
? Less critical, less frequently used computer files be relocated to data storage
facilities in the off-site facility, which could be retrieved on-line by the on-site
computers;
? A modified system architecture (computing, communications, and user serv-
ices) with its implementation plan be developed by ODP and OC to support the
relocation by 1992 of some computing services off-site;
? All future ODP budgets for procurement of large computer systems be
accompanied by communications plans from OC and facility plans from OL;
and,
? The user offices be required to budget for the start-up costs (computer
procurement, applications development, and facility expansion) for major new
ADP systems. (The Directorate for Administration would budget for common
user systems serving many offices.)
ES-2
UNCLASSIFIED
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
STAT
STAT
Rough budget estimates (highly site dependent, recurring costs not included, and no
addition for inflation) for implementing the above recommendations are as follows:
? Construction of a new off-site facility, including security and support
The Panel notes that locating the new remote facility on the Langley compound
rather than off-site could eliminate most of the communications costs and risks
associated with a remote operation. The Panel did not recommend the Langley
location for two reasons:
? Local community opposition might indefinitely delay its construction.
? The remote facility's benefits as a backup would be partially lost.
? Providing redundant, secure communications to remote site
ES-3
UNCLASSIFIED
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
REPORT OF THE CIA COMPUTER PANEL
ON CONTAINMENT OF ODP COMPUTER SPACE GROWTH
IN THE EXISTING AND NEW HEADQUARTERS BUILDINGS
STAT
STAT
1. INTRODUCTION
At the direction of the CIA Deputy Director for Administration, astudy panel was
convened on S November 1984, and given the charter to "determine the most feasible
and cost-effective method to prevent the continuing displacement of personnel in the
Headquarters building" (including the new Headquarters building addition) "by
machines." Membership on the Panel consisted of the following:
Panel's Executive Secretary, and a support staff consisting of representatives from
the Offices of Data Processing (ODP), Communications (OC), Logistics (OL), and
Security (OS).
Early in the Panel's deliberations, upon advice from the Deputy Director for
Administration, it was decided to focus on the machines (computers and associated
equipment) that are the responsibility of ODP. The Panel is aware that Agency
computer growth outside of ODP's purview could also result in displacement of
personnel. However, the Panel did not review non-ODP plans and is not offering any
comments in that area. Additionally, the requirements baseline for the Panel's study
was the ODP-provided projections of growth in machines and services, and ODP
plans for satisfying that growth. The Panel did not attempt to verify those projections
or ODP's plans.
The Panel requested and received a series of briefings and handout material from
ODP, OC, OL, and OS. These included information on:
a. The current ODP computer center and workstation environment, the services
it provides, its expected growth, and its transition to space that will be
available in the new Headquarters building addition.
b. Present and planned communications support of the ODP machine
environment.
c. Feasibility, costs, and risks of providing secure communications to a future
ODP environment which would include off-site computers and associated
files.
The Panel was assisted in its work by of CIA, serving as the
d. Options for splitting the ODP operations by service (some on-site and some
off-site).
e. Considerations involved in site acquisition and new building construction, on-
site and off-site.
f. Security considerations in providing ODP services from an off-site location.
1
UNCLASSIFIED
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
STAT
STAT
STAT
STAT
STAT
ODP currently has the following space allocated for large mainframe computer
centers in the present Headquarters buildin?:
TOTAL 47,462 sq. ft.
The new Headquarters building addition has 200,000 sq. ft. designed for accommo-
dating technical equipment (computers and communications). The utility support (air
conditioning, UPS, etc.) for the Headquarters compound is being upgraded to
accommodate a total of 200,000 sq. ft. of technical equipment space in the present
Headquarters building and the new Headquarters building addition. 50,000 sq. ft. of
machine space in the new Headquarters building addition is reserved for the Office of
Communications. ODP plans to use the remaining 150,000 sq. ft. Initially, 60,000 sq.
ft. of this will be used for machines, and the remaining 90,000 sq. ft. will be occupied
with ODP people. As machine space requirements grow, the ODP people will be
relocated to make room. When the new Headquarters building addition is ready for
occupancy in 1987, ODP plans w Center in the new Headquarters
building addition, release the and rearrange all the services as
follows:
At that time, according to ODP lans 96 000 s . ft. will be available for machines
(36,000 sq. ft. in the present plus a 60,000 sq. ft.
"New Center" in the new Headquarters building addition.) This will satisfy ODP
space needs until 1990. Beginning in 1990, further growth in space requirements
would be largely met by enlarging the New Center at the expense of ODP office
space in the new Head uarters building addition. However, some expansion of the
s planned by ODP beginning in FY 90. At this rate
of growth, the load on the utility support system would be nearing the limit by the
year 1995, with ODP using about three-fourths of the utility systems capacity, and
OC using one-fourth.
The Panel believes that it is not possible to put a ceiling on the growth in ODP serv-
ices and the Agency's eventual space requirements for information processing. The
behavior patterns of analysts and consumers of the intelligence product are changing
in response to the new capabilities available, and the volume of information to be
processed is continuing to grow explosively. Instead, measures must be sought which
will moderate the rate of growth in space requirements where feasible, and channel
the growth in a direction where it will not compete for Headquarters personnel space.
The balance of this report addresses these measures, which include technical,
managerial, and new building options.
2
UNCLASSIFIED
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
3. TECHNICAL OPTIONS
There are a number of technical options available to the Agency which may offer
some reduction in Headquarters space requirements for ODP computer systems.
These include:
? Use of new technology, direct access, data storage devices which provide
greater storage density per unit volume.
? Architectural changes which enable space consuming on-line computer files to
be remotely located.
? Use of new technology computers which provide greater processing capacity
per unit volume.
The Panel considered each of these options as discussed below:
3.1 Direct Access Storage Devices
Figure 1 shows the historic growth in space requirements for ODP computer systems
and the projected space needs through 1995. The space requirements in Figure 1 are
further broken down to show which part is for Mainframes (computers), Direct
Access Storage Devices (DASD) and All Other (tape drives, printers, etc.). As can be
clearly seen, the major contributor to the explosive growth in space requirements is
DASD. Figure 1 does not take into account any reductions in space needs that might
result from technological improvements.
STAT The Panel asked one of its members, to conduct an
assessment of what improvements in ata storage density might be possible as the re-
sult of technical improvement. ODP also provided estimates of possible space
reductions from technical advances. The results of these studies are summarized in
Figure 2. This provides a snapshot of what might be expected in 1995-10 years from
now.
STAT
51H1
STAT
ODP estimates that 17,146 gigabytes (gb) of storage capacity must be available in
1995. Assuming that space available to ODP for computers will be limited to a fixed
amount (100,000 sq. ft.), only 3,247 gb could be provided in this limited space using
today's technology. ODP's planning is based on a 4X improvement in storage density
by 1995 due to technical improvements. This would provide a capacity of 12,990 gb
in 1995 which is still inadequate to satisfy the requirement of 17,146 gb. The
study projects a minimum capacity of 17,800 gb, about adequate to satisfy the
need. This estimate is based on the most advanced magnetic storage technology
expected to be available with some optical storage.0 also believes that a
predominately optical storage capability might be available in 1995, thus providing a
capacity as high as 71,200 gb. On the other hand, ODP's estimate of the maximum
capability that technology might provide is 44,750 gb.
ODP's plans reflect a continuing improvement in storage capacity due to technical
improvements as shown in Figure 3. But they believe that the rate of growth in re-
quirements will exceed the rate of growth in capacity and that requirements will
exceed capacity by 1992 if space is limited.
The Panel has concluded that substantial improvements in storage technology can be
expected in the future; however, it cannot be certain that the improvements will be
sufficient to satisfy growing needs if space continues to be limited. As can be seen in
3
UNCLASSIFIED
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
}>Py artcnbg
4
UNCLASSIFIED
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
.~ .....................a........................~. ~......
5
UNCLASSIFIED
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
i 8~~ 1~ i f 1 ~g
~8cra~ yo ~~~~1J
6
UNCLASSIFIED
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Figure 2, there is a 5 to 1 spread in estimates regarding what technology might do for
us. Furthermore, the growth projections also introduce another element of uncertain-
ty. These projections reflect ODP's best judgement based on growth expected in
today's services. Major new systems-not identified at this time-could cause storage
growth even greater than that projected.
3.2 Remote Computer Files
Because of these uncertainties, the Panel believes that ODP must develop an
alternative to continuing to locate all on-line files (DASD and tape) in Headquarters
space. Architectural changes should be developed to permit the relocation of some of
these files to a remote site, but accessible on-line via secure, redundant communica-
tions. It is likely that some compromise in access time would need to be accepted by
the users to keep the costs of such an arrangement within reasonable bounds.
Therefore, files to be relocated would be those infrequently accessed and those which
would be less time-sensitive. The Panel believes that the uncertainties are such that
ODP could well reach the saturation point in DASD space in the and-90's. Therefore,
remote storage capability should be in place by that time.
3.3 New Technology Computers
In Figure 1, it can be seen that mainframe space requirements have been, and are ex-
petted to remain, relatively flat. Also, mainframes are a relatively small part of the
total space required. However, some space savings would be possible from greater use
of new technology mainframes than that presently planned by ODP. The new
technology machines would provide about twice the processing power as the largest
old machine in the same amount of floor space. ODP presently plans to acquire 15 of
the new technology machines between FY 85 and FY 91, but also plans the continued
use of 3081 and 3083 mainframes (old technology) to support some SAFE, DDO, and
common user services. Thirteen of these old technology mainframes could be replaced
by seven of the new technology machines, soon to be available (Siena machine). But,
even if this were done, it would release about 1,450 sq. ft. which is not nearly enough
to accommodate DASD space requirements. While there may be some operational
advantages from the new technology machines that were not addressed by the Panel,
it is clear that the space savings alone from greater use of these machines is
insufficient to significantly impact the space requirement for DASD.
7
UNCLASSIFIED
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
4. MANAGEMENT OPTIONS
For years, senior Agency management has struggled with the question of how to best
manage the growth of computer services in the Agency. During that time, computers
have continued to grow and have now become a powerful and essential tool enhancing
the Agency's operational productivity.
Historically, for a variety of reasons, the Agency has chosen to keep most of its ODP
computing resources in the Headquarters. When additional space was needed to
acxommodate computer growth, it was obtained at the expense of office space in the
Headquarters. Space for the displaced personnel, as well as for other personnel who
could not be fitted into the Headquarters, was obtained by converting dollar resources
into leased off-site o>~cx space. ('There is currently leased space in over 30 buildings).
Management has reviewed and approved the growth in on-site computing as
appropriate and essential to mission performance. Each new requirement was handled
separately on its individual merits, and there was little incentive or opportunity to de-
velop ageneral solution to the resultant space problem, i.e., the gradual displacement
of people to off-site locations.
In order to assure that future space requirements are given proper consideration as a
part of the overall resource picture when new systems are proposed, two things are
essential. First, the total resource implication of a proposal for a new (or significantly
expanded) computing requirement, including space, must be identified. All the
resources needed for establishing the new or significantly expanded capability should
be budgeted by the user directorate (Common user systems serving many users will be
budgeted by the Directorate for Administration, as will the continuing operational
costs of the systems). Second, a more responsive mechanism for converting the
budgeted funds into required new computer and support space must be established.
The required new space must not be in the Headquarters or its new addition,
otherwise people will continue to be displaced.
These new procedures, if adopted, will achieve two important goals. First, the user-
level organizations will participate in justifying and obtaining resources to support
new and expanding requirements. This should result in more careful screening of
needs when they are in competition for directorate resources. Second, management
has available another option for satisfying computer space needs other than personnel
displacement, namely, acquiring new space for expansion with budgeted dollars.
8
UNCLASSIFIED
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
From the preceding discussions it can be seen that some technical and managerial op-
tions could contribute to the solution of the space problem, and both point towards the
acquisition of additional computer grade space, not derived from the Headquarters or
its new addition, which can be readily expanded as future needs develop. In this Sec-
tion the Panel examines the ancillary benefits from establishing a separate computer
grade facility and optional approaches to its establishment.
5.1 Security, Backup, and Continuity-of-Service
Establishing a computer grade facility separate from the Headquarters and its new
addition would have the added benefit of providing backup to vital computer services
in the event of a major Headquarters catastrophe. Currently, continuity-of-service is
vulnerable to fire/water damage and possible damage from a terrorist attack.
Information stored in Agency computers is already backed-up onto magnetic tape
and stored off-site to enable reconstitution of vital files in the event of catastrophes
small and large. The use of a connected off-site facility to house the backup tapes,
and perhaps provide some on-line backup, can provide for speedier recovery. The
security implication of two computing sites vice one is principally the cost of
defending a second perimeter. It is unclear if a single large target is more or less
attractive to hostiles than either of two smaller ones. A coordinated attack on two sep-
arated, secure installations requires much more complicated planning and much more
careful execution. There are added costs in securing a second site to the same
standards as the current Headquarters facilities. Moreover, an off-site facility with
accommodations fora (small) user community could be used for emergency people re-
location as wcll.
Damage to habitable space in Headquarters could be compensated for by temporary
relocation of personnel to an off-site facility where they could be properly supported
by proximal computing.
5.2 Range of Options Considered
-Anew expandable, technical equipment building could be constructed on-site
or off-site;
-This building could accommodate either part or all of the Agency's computer
equipment;
-The division between this building and Headquarters could be either by type
of equipment (e.g., all DASD and controllers moved out of Headquarters) or
by function (e.g., the complete computer systems needed to support Commu-
nity information systems moved out of Headquarters or by mixes of
equipment and function).
The CIA support staff, at the request of the Panel, performed and reported on a num-
ber of studies to assist the Panel in understanding the feasibility, costs, risks, and
other considerations associated with these options. To bound the support staff studies,
the Panel identified representative off-site locations to be considered for studying
variations in costs and operations (construction costs, communications, security,
manning, etc.). The Panel also requested a study of an on-site location for the new
technical equipment building.
9
UNCLASSIFIED
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
The support staff also provided a set of six options for dividing the ODP workload
functionally. They are cumulative: the first (Option a) includes the services ODP
considers easiest to move; the second adds the next easiest group. The sixth includes
all ODP services. The options are:
a. Community systems are moved out, the rest remain at Headquarters.
b. The Community systems and those used in developing new software are
moved out, the rest remain at Headquarters.
c. The Community, development, and batch processing systems are moved out.
d. The Community, development, and batch processing systems are moved out,
along with the GIMS and VM services.
e. All the above, plus the Project SAFE services are moved out. Only the special
DDO services remain at Headquarters.
f. All data processing is moved out.
All six of the options appear to be technically possible, though the degree of risk in-
creases from (a) through (f).
The ODP staff prepared estimated space requirements for the six options of dividing
the workload functionally. They are summarized in Table 1. The first six columns
contain estimates for different parts of the computer facility including computer-
grade space. The next three are for support space. (The square footages mentioned in
the discussion below are for computer-grade space. Support space is not shown, but
can be found in Table 1.)
5.3 Reasons for the Panel's Selection
During the process of narrowing the options, the Panel made three decisions:
? The new technical equipment building should not be further away than nearby
Maryland or Virginia.
The costs and risks of being more distant were exemplified in the study of a
New England location; they are substantial. Choice of sites should take into
consideration possible benefits from existing plans for serving candidate sites
with high-speed communications.
? The division of computer equipment between the new building should be
primarily based on junctions rather than type of equipment.
While equipment at the two sites would be based on the functions supported,
each site could contain complete, interchangeable, computer systems, which
(with appropriate planning) could support the highest priority services of the
moment and provide mutual backup. Some tertiary or archival storage might
properly be put entirely in a new off-site building in the preferred area.
? A ceiling should be placed on the amount of computer space permitted at
Headquarters.
This would permanently solve the people dislocation problem, because by
definition, all growth beyond the ceiling would be in the new off-site building.
The question then is: how high should the ceiling be? The remainder of the
Panel's deliberations focused on this issue.
~o
UNCLASSIFIED
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
STAT
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
The Panel's final choice was among three options. Option 1 assumes that all computer
facilities will be moved to a new building (equivalent to ODP's functional division
Option f, in Table 1). Option 2 will set a ceiling of 50,000 sq. ft. of computer space at
Headquarters (initially corresponds to ODP's functional division Option c, and later
to Option d). Option 3 will set a ceiling of 100,000 sq. ft. of computer space at Head-
quarters (corresponds to ODP's Option c).
5.4 Description of Space Options
Option 1: Under this option, all ODP computer facilities would be moved out of the
Headquarters building and its new addition, thus requiring no space in these
buildings. Under one variation of Option 1, the new expandable technical equipment
building would be built on the Langley compound to house the ODP computers.
Under another variation, the new building would be built at a remote site.
It is inconceivable that this new off-site building could be ready for occupancy in time
to avoid some rearrangement of ODP computer facilities and services in 1987, when
the new Headquarters building addition becomes available. Considering budgetary
and construction lead times, availability of the new building in 1992 might be possible
to achieve. At that time, about 100,000 sq. ft. of computer-grade machine space
would be needed to accommodate the ODP equipment. Some upward adjustment in
size would be needed if initial availability is delayed beyond 1992. Cost of
construction at each of the 5 sites studied (as estimated by OL) ranged from
for the 100,000 sq. ft. computer-grade space and required support facilities. A
summary of these costs is included in Table 2 (see paragraph S.5 for communications
cost estimates). The two large mainframe computers that ODP plans to acquire in FY
89 and FY 90 could provide a vehicle for starting the transition of computer services
to the new building, when it is ready for occupancy in 1992.
This option has the main advantage of completely relieving the pressures to expand
ODP computer space in the Headquarters building at the expense of people. When
more space is needed, it will be obtained by incremental expansion of the on- or off-
site new building designed and dedicated to house machines. If an on-site location for
the new building is chosen, the added communications and security problem would be
minimal. On the other hand, an off-site location would have significant communica-
tions costs and risks. Redundant, high capacity, secure communications links would
be needed to tie Headquarters and the on-line users to the remote computer facility.
OC estimates that redundant, high capacity communications s stems to satisfy this
requirement could involve one-time costs ranging from depending
upon the location of the remote site. These links would be vulnerable to disruption
either intentionally or accidentally. The bandwidths required may push the communi-
cation state-of--the-art which, in turn, could limit the architectural alternatives which
might be considered for the terminal network.
The Agency has invested substantial resources in the construction of computer grade
space and supporting utilities in the present Headquarters building. Additional
resource expenditures are planned to prepare the machine space in the new
Headquarters addition, and to upgrade the Langley utility support. These invest-
ments would be lost under Option 1, except that some use could be made of upgraded
utilities if the new technical equipment building were built on-site at Langley.
12
UNCLASSIFIED
STAT
STAT
STAT
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
COST ESTIMATE, STAND-ALONE COMPUTER BUILDING,
AND SUPPORT FACILITIES WITH 100,000 SQ. FT.
OF COMPUTER-GRADE SPACE
SF gW TON
;MIL
Computer Floor
100,000 4,500 1,250
37.88
Office Support
38,500 577.5 80.2
5.775
Facilities Support
34,200 102.6 14.25
6.84
Power Plant
20,000 1,840 8.33
13.2
Warehouse
7,500 15 3.13
1.125
Incinerator
- 100 -
.75
Site
(50 acres) - -
.75
Security'
- - -
.689
TOTAL
200,200 7,135.1 1,355.91
67.009
Total Facility Costs by Site"
Coat, Million ; (M)
'RFI shielding coats are incorporated in facility construction costs.
"Includes variation on above total of ;67.009M, to take into account peculiarities of each of the five
sites. No addition for inflation.
13
CONFIDENTIAL
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Option l: Under this option, a fixed amount of space of 50,000 sq. ft. in the
Headquarters and new addition would be allocated for ODP computers. The
remaining ODP computers would be installed at a remote off-campus site and tied to
the Headquarters and other users by redundant, secure, high capacity communica-
tions links. The services that would continue to be offered out of the Headquarters-
located computers would be those whose users generate high communications
volumes, and whose services would be severely degraded by communications prob-
lems. The less communications-intensive services would be provided from facilities
located in a new building constructed at a remote site similar to that in Option 1. The
off-site facility should have about 50,000 sq. ft. of space available initially, and
probably would require the same budgetary and construction lead time as in Option 1
(availability no earlier than 1992). (It is noted that the new 50,000 sq. ft. building
could be built on-campus as previously discussed under Option 1. However, the Panel
does not support this approach because of the uncertain political climate, as
previously stated.) For the sites considered, construction of the 50 000 sq. ft.
computer-grade space and support facilities range from Table 3
provides a summary of those costs. (See paragraph S.5 for communications cost
estimates.)
ODP could begin rearrangement of services within the Langley complex when the
new Headquarters building addition becomes available in 1987. Services to be moved
to the new off-site center could remain in place in the present Headquarters computer
centers until the new off-site center is ready for occupancy in 1992. However, if the
Headquarters space in the present computer centers is needed sooner for offices, the
services to be moved could be temporarily relocated to the new Headquarters building
addition. During this transition period, in either case, the 50,000 sq. ft. Headquarters
ceiling will be exceeded.
The Panel believes that the logical services to be moved off-site to the new center
under Option 2 would be the Community, development, and possibly batch services.
The 50,000 sq. ft. new off-site center would initially accommodate these services with
some growth margin. There is concern, however, that this option does not provide suf-
ficient flexibility to accommodate growth in the services that would remain on-site.
Other significant measures would need to be taken eventually to prevent Headquar-
ters machine growth from displacing people. These could include off-loading large
segments of infrequently used shared data files to the off-site location. This, of
course, could quickly consume the small growth margin planned for that location.
There is a significant advantage to be gained from splitting the ODP operations as
proposed in Option 2 and also in Option 3 which follows. Under this arrangement,
some portion of ODP's processing capability at the remote site is protected from
catastrophic loss caused by a disaster in the Headquarters building and vice versa.
Furthermore, with careful planning and some additional investment, the remote
facility could serve as a full backup to the Headquarters system. This would solve a
long-standing Agency problem with modest additional cost.
The communications risks and costs would be lessened under Option 2 because efforts
would be made to keep the communications-intensive ODP services on the Langley
compound near their users, where providing reliable communications is not a
problem. Security risks would also be lessened, but access to the remote processing
capability and files could be more readily denied because of dependence on long
communications links.
14
UNCLASSIFIED
STAT
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
COST ESTIMATE, STAND-ALONE COMPUTER BUILDING
AND SUPPORT FACILTITES WTfH 50,000 SQ. FT.
OF COMPUTER-GRADE SPACE
Computer Floor
50,000 2,250 625 1
9.3
Office Support
20,000 300 41.67
3.0
Facilitld SUpport
10,000 30 4.17
2.0
Power Plant
12000 924 5
7.254
Warehouse
2,500 5 3.125
.375
Incinerator
- 100 -
.75
Site
(30 acres) - -
.450
Security'
- - -
.563
TOTAL
94,500 3,609 678.965 3
3.692
Total Facility Costs by Site"
'RFI shielding costs are incorporated in facility construction costs.
"Includes variation on above total of i33.692M, to take into account peculiarities of each of the five
sites. No addition for inflation.
15
CONFIDENTIAL
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
As in the case of Option 1, the failure to fully utilize all the computer-grade space in
the Headquarters building and its addition would forego any benefits from that
investment.
The Panel favors Option 2 over Option 1, but is concerned that all of the vital com-
munications-intensive services could not be adequately provided from the Headquar-
ters compound beyond 1992 if the available space is limited to 50,000 sq. ft. If some
of these services or files had to be shifted to the remote site, communications costs
would rise, and reliability and responsiveness might suffer.
Option 3: Under Option 3, a fixed amount of space would be allocated to ODP in the
Headquarters complex. It is suggested that this allocation should be 100,000 sq.ft.
and that it be provided as follows. While it would be feasible to retain some or all of
the present Headquarters space as part of the 100,000 sq.ft. allocation, the Panel
believes that utilization of the machine space in the new Headquarters building
addition is to be preferred over continued occupancy of computer center space in the
present Headquarters. Initially, 60,000 sq.ft. in the new addition would be used for
computers with the remainder providing room for growth, but temporarily used for
ODP office space. The TEMPEST protection afforded by the new addition appears to
be consistent with the President's Directive on Automated Information Systems
Security (NSDD 145). Furthermore, the benefits from the investment made to
provide TEMPEST protection in this space would be lost if it was not used for
machines. The transition would begin in FY 1987, when space in the new
Headquarters building addition becomes available and newly acquired computers
begin to arrive.
Additionally, a new 50,000 sq.ft. expandable facility would be constructed off-site to
house ODP computers which would be connected to the Headquarters user popula-
tion by secure, redundant, high capacity communictions links. The estimated cost of
constructing the new omputer grade space and required support
facilities range from for each of the 5 sites considered. (See STAT
paragraph 5.5 for communications cost estimates.) The computers located at Langley
would be used to provide vital communications-intensive services which, if located
off-site, would be severely degraded by communications problems sometimes experi-
enced with long off-site links. Those located in the new off-site building would provide
services which are more tolerant of communications problems, such as Community,
development, and batch services. Under this Option, ODP would work towards
releasing the present Headquarters machine space of about 47,000 sq.ft., and
eventually occupying a maximum of 100,000 sq.ft. for machines in the new
Headquarters building addition.
Because the communications-intensive services would continue to be provided from
the Langley compound, communications costs and risks would be reduced considera-
bly when compared to off-site location of these services. However, those services
served by the remote facility would be more vulnerable to disruption because of
communications failures.
The Panel favors Option 3 over Options 1 and 2. This Option provides sufficient on-
site space to meet expected growth of those vital services whose users could best be
served by computers located in the Headquarters compound. There can be flexibility
built into the schedule for this Option to make allowances for delays in obtaining con-
struction funds for the off-site new facility. The allocated space in the new
16
UNCLASSIFIED
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Headquarters building addition can be used immediately, when available, to begin to
configure the equipment and services that will remain when the other equipment and
services move to the off-site new building.
5.5 Communications Costs
OC provided input to the Panel regarding cost and feasbility of providing communi-
cations support to the new dedicated technical equipment building, which could be re-
quired under all of the three options studied. Each of the candidate sites under study
was examined in the course of the study. Rather than attempt to analyze all possible
options, OC selected two cases~ne in which only Community Systems would be
served from the new off-site building, and the other in which all of the ODP on-line
systems would be served from the new building. Additionally, OC examined how the
costs and feasibility would be affected by two different communications architectures
now under consideration by OC-one known as the WPN (Wide Premises Network)
architecture and the other known as the PBX architecture. Tables 4 and 5 summarize
the results of those studies. For the WPN architecture, the one time costs for
establishing the communications network servin the 5 candidate sites can range up
~1~I with recurring a e costs~For the PBX architecture, the ~STAT
one time costs range up to flual-lea ith annual costs up to~ Communications STAT
cost and technical considerations favor the on-site (Langley) location or, as a second
choice, those off-site locations in Northern Virginia.
17
UNCLASSIFIED
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
25X1
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
6. FINDINGS AND RECOMMENDATIONS
6.1 Findings
? ODP computer equipment now occupies less than 50,000 sq. ft. of computer
space at Headquarters, but significantly more will be needed soon. The
150,000 sq. ft. of computer-grade space which will be available to ODP in the
new Headquarters building addition in 1987 could potentially satisfy all ODP-
estimated growth requirements up to the year 1995. However, to accommodate
its machines in 1987, ODP plans to retain 36,000 sq. ft. in the present
Headquarters, plus using 60,000 sq. ft. in the new addition, with the remainder
used for ODP offices initially. The majority of future growth will be
accommodated by displacing ODP personnel in the new addition, who are
occupying the unused computer-grade space. By the year 1997, the entire
150,000 sq. ft. of the new addition will be required for computers and data
storage, in addition to the 36,000 sq. ft. of present Headquarters space which
would be retained by ODP. Before that condition is reached, the utility support
capacity available for technical equipment in the Headquarters complex will
be over committed, and no further expansion of on-site computer systems
would be possible without increasing the capacity of utility support.
? Special features (such as TEMPEST protection) were planned and are being
designed into the new Headquarters building addition to accommodate
technical equipment which processes classified information. The Panel believes
that this space should be utilized for its intended purpose, and that the 36,000
sq. ft. in the present Headquarters computer centers be released when the new
Headquarters building addition space becomes available.
? Aggressive pursuit of new technology will help postpone space saturation
problems in the Headquarters and its new addition. The greatest benefit will
be realized in the Agency's fastest growing area of space consumption~lirect
access storage devices. Technology improvements, such as double-density
magnetic disks and optical disks, are expected to yield as much as a 4X-9X im-
provement in storage density by 1995. Additionally, new technology main-
frame computers will produce twice as much work in the same physical volume
as old technology. However, the Panel concludes that technology improve-
ments alone will not produce sufficient space savings to solve the space
problem.
? As long as the computers and people were competing for the same space, and
the Agency was committed in its architectural approach to locating all its
computers in the Headquarters, there was no alternative but displacement of
people to accommodate computer growth. The displaced people could be
readily accommodated elsewhere, as is evidenced by the 30+ off-campus
locations for offices.
? While there may have been valid reasons in the past for not doing so, today it is
technically possible to relocate all or part of ODP's processing capability to a
remote site. Staff studies performed at the Panel's request indicate that
security, communications, and operational problems are all manageable.
However, it seems prudent to retain the most communications-intensive
services at Headquarters (little would be gained by moving them all out, and
substantial added cost and risk would be incurred). As part of the relocation of
19
UNCLASSIFIED
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
some of the Headquarters computer files to the remote site with on-line access
from the Headquarters computers. The Panel also believes that if some
computers and files relocated to a new, separate building, then it should be lo-
cated at any one of several possible sites in nearby Maryland or Virginia.
? It is important to note that a division of ODP services will have a major impact
on the ADP system architecture, but not to the extent that technical risks are
unreasonable. There is a significant peripheral benefit from a divided opera-
tion. With proper planning and design, the remote facility and the Headquar-
ters facility can back each other up, providing protection against a catastrophe
at either site.
6.2 Recommendations
The Panel recommends that:
? The Agency set a ceiling of 100,000 sq. ft. of space for ODP computers in the
Headquarters compound, which would be obtained from the computer-grade
space in the new Headquarters building addition;
? The 47,462 sq. ft. of computer space in the present Headquarters building be
released for use as office space;
? An expandable, computer-grade facility of initially 50,000 sq. ft. plus support
space be provided (purchased or leased) off-site by 1992 for ODP computers
and peripheral equipment;
? The off-site facility be located at any one of several possible sites in nearby
Maryland or Virginia;
? The off-site facility be used to provide services which are less communications-
intensive (i.e., Community, development, and possibly batch applications);
? The off-site facility be linked to the Headquarters building via redundant,
secure, high bandwidth communications;
? Both the off-site facility and the Headquarters building contain computer
systems sufficiently similar to each other to provide mutual backup;
? Less critical, less frequently used computer files be relocated to data storage
facilities in the off-site facility, which could be retrieved on-line by the on-site
computers;
? A modified system architecture (computing, communications, and user serv-
ices) with its implementation plan be developed by ODP and OC to support the
relocation by 1992 of some computing services off-site;
? All future ODP budgets for procurement of large computer systems be
accompanied by communications plans from OC and facility plans from OL;
and,
? The user offices be required to budget for the start-up costs (computer
procurement, applications development, and facility expansion) for major new
ADP systems. (The Directorate for Administration would budget for common
user systems serving many offices.)
20
UNCLASSIFfED
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
u~ u i i
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2
CONFIDENTIAL
CONFIDENTIAL
Declassified in Part -Sanitized Copy Approved for Release 2013/01/02 :CIA-RDP87M00539R000901160019-2