|
| |
From the web site at:
http://www.pirp.harvard.edu/pubs_pdf/stenbit%5Cstenbit-i03-1.pdf
-----------------------------------------------------------------
A CONVERSATION WITH THE ASSISTANT SECRETARY OF DEFENSE FOR C3I
John P. Stenbit
March 13, 2003
John P. Stenbit became the assistant secretary of defense for command,
control, communications, and intelligence (ASD(C3I)) in August 2001, and
now heads the C3I successor organization, Networks and Information
Integration. Prior to this appointment, his private and public sector
service in the telecommunications and command and control (C2) fields
spanned over thirty years. From 1973 to 1975, he served in the Department
of Defense (DOD) as principal deputy director of telecommunications
and C2 systems; for the next two years, he was staff specialist for
worldwide military command and control systems in the Office of the
Secretary of Defense (OSD). He has served as chairman of the Science and
Technology Advisory Panel to the director of central intelligence (DCI),
and was a member of the Science Advisory Group to the directors of naval
intelligence and the Defense Communications Agency. He also chaired the
Research, Engineering and Development Advisory Committee for the
administrator of the Federal Aviation Administration, and has served on
the Defense Science Board, the Navy Studies Board, and the National
Research Council Manufacturing Board. In 1968 he joined TRW and was
responsible for the planning and analysis of advanced satellite surveillance
systems; he retired from TRW as an executive vice president in 2001.
Previously, he had worked for the Aerospace Corporation on C2 systems
for missiles and satellites, and on satellite data compression and pattern
recognition. Mr. Stenbit has bachelor's and master's degrees in electrical
engineering from the California Institute of Technology, and is a member
of the National Academy of Engineering.
-------------------------------------------------------------------------
Oettinger: We are privileged to have as our guest today John
Stenbit. You have had a chance to look at his biography and find
out about his distinguished career, so I won't eat into his time
by making any further introductions. He has indicated a
willingness-indeed, a desire-to be bombarded with questions
rather than to give any kind of formal presentation. I know
you're not shy, but just to avoid the slightest risk that we
won't get off to a start, let me throw the first one out. Could
you enlighten us a bit about what pulled your office
together-what the motivations were for pulling together command,
control, and this, that, and the other thing; what the
motivations are currently for creating another office with the
intelligence label; and where that all stands practically and
rationally? Let's go around. Other questions?
-2-
Student: We discussed moving from n 2 communications nodes to n.
Could you talk a little bit about the bandwidth associated with
that?
Student: I was wondering if you could talk about how the
"I" part of your organization is changing as a result of the war
on terrorism. What happened to "I" after September 11?
Student: I'm curious about funding. Do you think that you're
properly funded? What role does the monetary system play?
Student: Could you maybe discuss the CONOPS [concept of
operations] for data management and information sharing? Then
I have a second question on the budget. Could you maybe discuss
your thoughts on Joint Forces Command [JFCOM] and how the PPBS
[planning, programming, and budgeting system] cycle might be
reformed to better incorporate joint technologies?
Student: What do you see as the changes in your office as you
go from "C3I" to "C3"?
Student: Do you have anything to say about the relationship
between the DOD and the new Department of Homeland Defense
[DHS], and if your agency is going to be involved with that in
any way?
Stenbit: Is this off the record? I have no problem with
anyone down here.
Oettinger: Let me spell that out. Anything you say is not for
attribution if you don't want it to be; however, we are recording
it. You will have a chance to edit it before anything gets
published, and we've had no leaks of the tapes in the past.
Student: How is the DOD involved in the Terrorist
Threat Integration Center [TTIC]?
Stenbit: I must admit I'm not an expert on that, so I think
that's going to be a short answer.
Student: Rob Slade just published a damning indictment of our
national strategy to defend cyberspace in RISKS DIGEST.(1) The
policy is supposed to make cyberspace safe. I wonder if you have
anything to do with this national strategy and what you think
about it.
Stenbit: I assume the article attacked the policy that
came out of the White House about four months ago-and the final
version that just came out? Information assurance policy is what
you're after here.
Okay, that's a good group of questions. I'll see if I can
get through those in forty-five minutes or so. That'll leave
enough time for additional questions.
Motivations and what happens when you go back and forth are
interesting. Whether it's a "C3" office or an "I" office or a
combination of the two, nobody in the OSD wanted either of them
to be there, because they symbolize a set of problems that always
come up. It would be
---------------
(1) Rob Slade, "Education and the National Strategy to Secure
Cyberspace," The Risks Digest 22, 63, 12 March 2003, [On-line].
URL: http://catless.ncl.ac.uk/Risks/22.63.html#subj1 (Accessed
22 July 2003.)
- 3 -
nicer if it were some sort of clean organization. For instance,
when you talk about financing there's a comptroller. When you
talk about personnel, there's a personnel office. Because it's
the government and the DOD, there is always going to be a policy
organization that's looking at what the State Department (or
whoever) does from our point of view. Because the DOD spends an
enormous amount of money, there's going to be some integration
of how we procure and acquire things. Those functions are
accepted as parts of the job, and have an evolution of their own.
They become stovepipes. They become an entire financial mafia
and an acquisition mafia and their attributes all get written
into the law.
Both "C3" and "I," in whatever form you'd like to take them,
are there because of a different kind of problem. Something went
wrong. Something didn't work, and it didn't work often enough and
consistently enough that we had to have somebody to task with
working on solving that problem. In the period of the 1960s, neither
office existed, as I recall. (It was slightly before my time.) In
the early 1970s, there was created what was called an ASD(T), an
assistant secretary of defense for telecommunications, and there
was an ASD(I), an assistant secretary of defense for
intelligence. They were parallel to each other, and they were
both there for approximately the same reason, which was the
demand for intelligence information from the operational side of
the military. That meant the information had to be very specific
and very fast, as opposed to (and I'm not trying to be
pejorative when I say this, but I'm trying to make an allusion
that is at least useful for you) sitting and reading things and
then thinking about them and writing reports about what you
read. The things you have as input are a little bit more
privileged as pieces of information than you can get by reading
books or newspapers. The intelligence community was really used
to large-scale analysis kinds of issues, but the output was in
the form of papers and almost books, et cetera.
The interaction of various force components and what was
going on in Europe in those days caused people to start to worry
and ask, "Hey, wait a minute! What are the details of the Russian
plans for attack? More important, what are the sorts of things
that we could measure to get some inferential data that might
actually lead to where people would go and attack things at a
tactical level as opposed to the strategic level?" I wasn't
there, and I can't speak for it, but my sense is that this
happened when we first believed that we were actually going to
do something other than drop nukes on somebody.
Oettinger: For those of you who are interested in pursuing this
topic further, there is a contribution to the seminar by Ruth
Davis, who I think was director of defense research and engineering
around that time.(2)
Stenbit: She came after this. She was there in the late 1970s,
after I left. She was at the DIA [Defense Intelligence Agency]
when all that was happening.
Oettinger: Second, we will be hearing in April from Jim Simon,
who would represent a somewhat different view of the balance
between intelligence in the large versus the tactical.(3)
-------------------
(2) Ruth M. Davis, "Putting C 3 I Development in a Strategic
and Operational Context," in Seminar on Command, Control,
Communications, and Intelligence, Guest Presentations, Spring
1988 (Cambridge, Mass.: Harvard University Program on Information
Resources Policy, I-89-1, March 1989), [On-line]. URL:
http://www.pirp.harvard.edu/pubs_pdf/davis\davis-i89-1.pdf.
- 4 -
Stenbit: He's younger than I am, so his first-hand data probably
also come from back in the 1970s.
Anyway, on the telecommunications side I do know what was
happening. We were having a whole series of major failures. The
North Koreans captured the USS PUEBLO in January 1963. [sic] It
took them thirty hours to get the PUEBLO back to North Korea, but
it took us thirty-six hours to find out that there were Marine
aircraft on Okinawa that could have intercepted it and saved it.
The fact that the Koreans were slow is interesting; the fact that
we were slower is even more interesting.
The Israelis called us up one day and said, "If you don't
get that ship, the LIBERTY, out of this place we're going to sink
it in twenty-four hours."(4) We couldn't tell the ship to move
when we got the data back because it was already under the water,
because it took more than twenty-four hours for the data to wander
in through the system and come out at the other end.
A ton of such case were going on, and there was a whole
committee in Congress carrying out permanent investigations of the
failures of Pentagon communications. The DOD telecommunications
and intelligence offices were formed at about the same time, and
they were both formed because of these problems. Actually, the key
for both of them was to get involved in the integration of
acquisition, development, policy, and actual strategy. We were able
to use computers -- the environment was getting more automated --
but in those days there weren't any networks. DARPA [Defense
Advanced Research Projects Agency] was just thinking about inventing
the ARPANET in 1969. I think Mr. Gore got caught up in this a little
bit, but what he did was about ten years later. It was basically
a 1980 phenomenon-before there was any rigorous and robust
network.
On the intelligence side it was the same issue. It was
highly compartmentalized. The NSA [National Security Agency]
didn't tell anybody what it did; the CIA [Central Intelligence
Agency] didn't tell anybody what it did. We were taking pictures
in those days, and the people who took them didn't tell anybody
what they did, and so forth. The genesis was sort of "Let me
have some places to go and work on this problem." What really
happened on the "I" side in 1974, 1975, and 1976 was a whole
stretch of congressional activity dealing with the intelligence
agencies, undertaken by Senator Frank Church [Dem.-Id.] and
Congressman Otis Pike [Dem.-N.Y.]. Some of us were suspicious
about what their motives were, but in any case they were
certainly effective in challenging a lot of
--------------
(3) James M. Simon was assistant director of central intelli-
gence for administration. See James M. Simon, "Crucified on a
Cross of Goldwater-Nichols," in Seminar on Intelligence,
Command, and Control, Guest Presentations, Spring 2001
(Cambridge, Mass.: Harvard University Program on Information
Resources Policy, I-01-3, July 2001), [On-line]. URL:
http://www.pirp.harvard.edu/pubs_pdf/simon\simon-i01-3.pdf,
and "Analysis, Analysts, and Their Role in Government and
Intelligence," in Seminar on Intelligence, Command, and Control,
Guest Presentations, Spring 2003 (Cambridge, Mass.: Harvard
University Program on Information Resources Policy, I-03-1, July
2003), [On-line]. URL:
http://www.pirp.harvard.edu/pubs_pdf/simon\simon-i03-1.pdf
(4) For a detailed account, see A. Jay Cristol, "The Liberty
Incident," in Seminar on Intelligence, Command, and Control,
Guest Presentations, Spring 1995 (Cambridge, Mass.: Harvard
University Program on Information Resources Policy, I-96-2,
January 1996), [On-line]. URL: http://www.pirp.harvard.edu/
pubs_pdf/cristol\cristol-i96-2.pdf and The 1967 Israeli Attack
on the U.S. Navy Spy Ship (Washington, D.C.: Brassey's Inc.,
2002).
- 5 -
issues about the intelligence community. They really went on a
crusade to challenge the removal of rights that had allowed the
intelligence communities to do certain things, compared with
normal systems. You have to remember that that's when Mr.
Rumsfeld was the secretary of defense. So when he thinks about
intelligence, what he remembers is having to hire ex-Congressman
[Robert F.] Ellsworth as another deputy secretary of defense (in
those days we had two deputy secretaries of defense, the only
time it's been like that), and he had Al Hall as the ASD(I).
They spent 100 percent of their time worrying about Church and
Pike and what we were going to do and how we were going to do
it. It became an enormous drain on the Secretary's time, in
addition dealing with all the problems that had caused the
ASD(I) to be formed in the first place. We used to refer to that
as the "Green Door Syndrome." The guys behind the green door sat
there and fondled the data and never told anyone anything about
it, et cetera. There was a lot of pressure to try to break down
some of those barriers and see if we couldn't use the
information more proactively.
On the C3 side, we had started enormously complex programs
to make radios interoperate with each other. We wanted to
communicate using secure voice. In those days, everything was
analog. There were no digital systems, so that meant you had to
digitize it. Analog-to-digital converters were rather complex
in those days. Then, after you digitized it, you didn't have
the bandwidth to send it, so you had to compress it. As soon
as you compressed it, everybody sounded like Donald Duck. In
fact, if you get a very low-bandwidth digital signal today
and you try to compress it, the vocoders are much better because
the processing gain is much better, but we didn't have any
processing then either.
So the bottom line was that we had to come up with a
standard that was DOD-wide, which was 16-kilobit voice, Vinson
encryptors, and vocoders, et cetera. Let me assure you, every
service already had its own systems, and none of them used
Vinson and none of them was 16 kilobits. So there was this
first-order problem about how to get the Air Force to talk to
the Army, et cetera. Those were very real problems, because the
DOD ended up with systems they developed themselves to special-
purpose specifications. There were no commercial standards for
any of this stuff.
Today interoperability is a lot easier. We may not know
whether we want to do GSM [Global System Mobile] or CDMA
[code division multiple access], but we can build a phone with
both of them in it. We didn't have any of those options then.
The phone would have been the size of a refrigerator if we had
put both in. We can do TCP/IP [Transmission Control
Protocol/Internet Protocol] and anybody can come with any
computer and hook it up to the Internet. There were no such
systems then. Ma Bell, in those days, would not allow you to buy
a handset and hook into the telephone system. It was a "foreign
attachment," and it was against the law to attach a handset to
the Bell network because it might have a resistance load that
might feed into something that might cause a relay to singe once
every 800 years, but they had convinced some court that was a
problem. So we are not talking about today's world. We're talk-
ing about everybody's stuff being proprietary and done by them-
selves. It was either done by Ma Bell or by the DOD, and so on.
Then other things happened. The DOD decided that it was
cheaper to buy computers in bunches rather than per application,
because computer hardware was expensive in those days. What
happens when you do a fixed-price, low-bid competition is that
you get the dog of the week. So we ended up with eighty red
Honeywell mainframes. That contract was the only thing
-----------------
- 6 -
that kept Honeywell in business. We were their last bid. GE
[General Electric] went out of the computer business while we
were in this time frame. All kinds of people went out of the
business. We managed to buy Honeywell. Their operating system
had nothing to do with IBM. Their disk drives had nothing to do
with anybody else. We got what we deserved. We got a really
cheap set of hardware that was useless. But, having bought it,
we now set up an entire program called the Worldwide Military
Command and Control System information system to build the
software to allow these dogs actually to perform a useful
function. That's the kind of job that we used to have in ASD(T).
It changed its name, but that's a trivial issue.
I was there for the last four years before it reorganized,
so I watched Mr. [James R.] Schlesinger's term as secretary and
during Mr. Rumsfeld's term as secretary. Neither one wanted us
there. When I would go up and see Mr. Clements, who was the deputy
secretary, and was the guy who actually made the decisions, he
would look at us coming and say "I'm going to get in trouble.
Every time you come in here, somebody screams at me right
afterwards."
We were discovering the National Environmental Protection
Act, which required the government to produce environmental
impact statements. We were trying to prove that if we modulated
a wire in Wisconsin the moss would not grow on the south side of
the trees instead of the north side and confuse the migrating
birds. Sixty Minutes was all over us every month.
Oettinger: I think it was a sizable chunk of Wisconsin real
estate!
Stenbit: No, it was a very small piece. It was about six
inches wide and a hundred miles long. In any case, the first
time we turned on one of those systems, they rang every
telephone in West Virginia, because they used the old rotary
dial telephones and every time you got any kind of electric
field they would ring. So there were some very interesting
problems going on.
We did some very useful things through interoperability.
In 1975, for the first time nuclear weapons owned by the Army
and the Air Force were fitted with devices that voided their
ability to be used unless they had external information. Prior
to that time, it was a highly intricate procedural set of
controls, but it basically was up to the folks who had them in
their hands. I think of that as an enormous stride forward in
the balance between "minimizing maximum regret" and still
being able to get something done if we needed to. How could we
do that? Because we had confidence in the coding, and confidence
in some communications systems that actually went worldwide
and were survivable, and we converted systems that needed fifty-
two aircraft working simultaneously to make them work to needing
three, so we saved money and made communications more reliable
by changing frequencies and changing the way we communicated.
So there was a lot of integration, if you will, that went
on in this particular timeframe, but it was all around special-
purpose stuff. Everyone could see that the end of that game was
already there: that we were going to be dominated by the
commercial world. Even in those days it became pretty obvious
what was going to happen. The Carter Administration decided to
change how this worked. President Carter said, "All of this is
just part of acquisition." That was a mistake, because he missed
the policy aspects of that. He took a gentleman named Gerald
Dinneen and he said, "You are now "C3" and "I," but you're also
the deputy acquisition guy in the department." So he took what
were fundamentally mixtures of policy, technology, and acquisition
and made them
---------------
- 7 -
into an acquisition-focused system, and lost the bubble a bit
(in my mind) about some of the more interesting interactions
that occurred in those days.
Oettinger: There is an account by Dinneen in the proceedings
of the seminar if you want to follow up on that.(5)
Stenbit: I'm not trying to say he did up, down, or indifferent.
He was assigned a job, like everybody else.
In any case, that lasted for four years. Then C3I kind of
was moved out from under acquisition, and has stayed there ever
since. The structure has to do with the policy, and it has to do
with strengthening the hand of the secretary to enforce things
such as interoperability-to actually stop people from doing
things that don't work with other people, et cetera. It's been
like that since 1981, so that's twenty-two years. Now Mr.
Rumsfeld is back, and he remembers it used to be split, and so
he just split it this week. "C3" and "I" have now split.
Oettinger: Is Steve Cambone confirmed as under secretary for
intelligence?(6)
Stenbit: He is confirmed as of this morning. All the jobs
involving integration of information are still present,
and therefore the argument to have an end-to-end purview of
what's happening with regard to information sources and
information flows is still a very valid one within the Pentagon.
But there's another problem, and that is the alacrity
with which the intelligence community serves what is in fact
an ever-expanding plethora of requirements from the Defense
Department for speed and accuracy and persistence, which are in
fact not consistent with sitting in a room doing studies and
writing reports. Once again, I'm not trying to be pejorative,
but you need to understand that there's a difference. Political
intelligence is not, in general, time sensitive. It's more like
history. It is predictive. Its utility is to help people
understand why other people are doing what they're doing. If you
happen to find out exactly what the position of the French
ambassador to the United Nations is before the vote next week,
it could be tactical, but what good does it do you? If you're a
company and you find out what the other guy is going to bid,
that's a different issue. You change your bid. So maybe it makes
a difference, but it is in fact fundamentally a slower process.
The bottom line here is that there's an enormous pressure to
increase the flow of information from the intelligence community
to the DOD over paths that haven't been used in the past. One of
them is to the secretary, by the way. We have a very interesting
system. There are committees that control the priorities within
the intelligence community, and on those committees sit the
secretary of state or his designee, the secretary of the
treasury or his designee, the national security advisor, an Army
person, a Navy person, and an Air Force person, but there's
nobody there for the Secretary of Defense. So actually, there is
physically no voice for the Secretary in the priorities of what
gets done. He doesn't like that, to put it mildly. There's a
whole raft of
------------------
(5) Gerald P. Dinneen, "C 3 Priorities," in Seminar on Command,
Control, Communications, and Intelligence, Guest Presentations,
Spring 1982 (Cambridge, Mass.: Harvard University Program on
Information Resources Policy, I-82-3, December 1982), [On-line].
URL: http://www.pirp.harvard.edu/pubs_pdf/Dinneen\Dinneen-i82-3.
pdf
(6) Dr. Steven Cambone was previously principal deputy under
secretary of defense for policy.
- 8 -
issues about intelligence information flow that Mr.
Rumsfeld is worried about, and he believes it's better to
concentrate on that and take a hit on the interaction and
the interconnectivity than it is to emphasize the inter-
connectivity.
He clearly thinks the interconnectivity is still import-
tant, because he wants the remaining part of whatever (we
haven't figured out what my office is going to be called)(7)
to emphasize end-to- end information integration, from
intelligence all the way up. It's more like a chief information
officer role, if you wish, than it is a special-purpose kind
of role. So that's the evolution.
The terrorism aspect that you brought up made that all
the more complex, because now the interface has to occur with
the law enforcement community as well as with the national
intelligence community, and that's hard. Those of you who have
not thought that through may think that everybody in the govern-
ment just fights with everybody else, so they don't tell any-
thing. You have to understand: we have a Constitutional problem
in our country about the difference between intelligence for
national security purposes and law enforcement. They're very
different. In the one case, the sanctity of grand jury investi-
gations and the secrecy of the information are in fact Consti-
tutional in nature and the information is not supposed to be
shared.
They have some other problems. When prosecutors bring
somebody to trial, their information is all subject to dis-
covery by the opponent, so it behooves them not to keep lots of
records of alternative hypotheses of the crime. In fact, it
behooves him to come to a conclusion quickly about the hypot-
hesis and then build a database that confirms it. That's the
opposite of intelligence, where you're trying to search for
alternative hypotheses. So there are institutional issues that
are very real.
On our side, we are very reluctant to hand information to
a law enforcement process where discovery has to do with chain
of evidence. The chain of evidence means the lawyers for the
defense can go back and find out where we got it. In a lot of
cases, we're not interested in telling an open court where we
got the information. We have secret systems. We have security
systems we built up to attempt to keep some of that secret. So
there are real problems that are not computer problems or
egomania problems or interoperability problems, but
Constitutional in nature, that cause sharing information to be a
very tough job.
Oettinger: Before you go on, one other footnote, because adapt-
ation to the issues that John has just mentioned is taking place
in a very dynamic sort of way. One of the longest histories of
accommodating to these issues is the creation by the Foreign
Intelligence Surveillance Act [FISA] of secret courts that have
a role of applying Constitutional principles, but in a closed
setting, so as to maintain secrecy. If some of you get involved
in this area, looking up the FISA court is one manifestation,
prior to the more recent enactment of the Patriot Act following
9/11, of adaptations, for better or worse, to these shifting
concerns.
Stenbit: That court is set up to protect the Constitutional
issues that I just mentioned. You get into the angels on the
head of a pin about "If I'm going to go do something, I have to
prove it has a foreign intelligence value." If I say it has a
law enforcement value, it can't go to the FISA court;
---------------
(7) On May 8, 2003, the assistant secretary of defense for
command, control, communications, and intelligence became the
assistant secretary of defense for networks and information
integration/DOD chief information officer.
- 9 -
it goes to a whole different court, with a whole different set
of procedures. Once I do that, and they give the approval, I am
then forced to live under the constraints that come with that,
which are severe, about "no American entities" and so forth. So
there's a rather intricate set of laws around what I was just
talking about.
Student: Who's winning the battles? The civil liberties advocates
or the security advocates?
Stenbit: I don't think it's a battle. We're not going to give up
our Constitution, so we're going to have a major challenge in this
world of bridging how we deal with the subject. It's not an
issue in my mind of either side winning; there are just some
fundamentals that are really hard.
Student: For example, with regard to facial recognition, the
security folks say that we want the cameras to recognize folks so
we can see who the terrorists are, and the civil liberties people
are saying "You're spying on us."
Stenbit: That's where I was headed. Let me head as close as
possible to the law enforcement versus foreign intelligence issue,
which is an interesting one. The addition of terrorism forces
us to face those issues both inside
and outside the United States. But the thing that's really a big
deal is that the DOD is not supposed to worry about anything
inside the United States. There is a constraint there, namely
the Posse Comitatus Act of 1878, and you should be happy that
there is such a constraint. It is in fact the DOD's job to
protect us from outside threats.
Oettinger: There's an article, if I could put my hands on it,
by Stewart Baker on the Posse Comitatus Act. It takes a somewhat
different view: that there's ample authority and precedent for
military intervention in a variety of domestic affairs that
included, for example, integrating the schools.
Stenbit: That was the National Guard. I don't think those were
ever active duty military.
Student: A previous speaker from Northern Command told us, "Yes,
there is this posse comitatus, but there are enough exceptions
that we're not too worried about it."(8)
Student: He actually said that troops were used in the inte-
gration situation and in the Los Angeles riots.
Stenbit: All of these things have waiver rights, and they
usually require presidential findings of some type. In fact,
almost all of these kinds of issues do have waivers that are
allowed. I'm not sure of the exact facts, but it's certainly
true that they would try to use the National Guard before they
would use federal troops, because of the posse comitatus issue.
But at the fundamental level, the DOD operates against
foreign threats, and the DHS is supposed to work domestically.
The DHS job is a lot harder than ours, because we clearly
have a
-----------------
(8) See Dale W. Meyerrose, "Adapting the Military to the Home-
land Defense and Homeland Security Missions," in Seminar on
Intelligence, Command, and Control, Guest Presentations, Spring
2003 (Cambridge, Mass.: Harvard University Program on
Information Resources Policy, I-03-1, May 2003), [On-line]. URL:
http://www.pirp.harvard.edu/pubs_pdf/meyerro\meyerro-i03-1.pdf.
-10-
decentralized security system within the country, what with
all the police and fire and various special services and so on.
That's not something that's ever been pulled together
particularly well, and that is a big job.
The one thing that we would like not to happen is what
happened after September 11. I'm not complaining, and I think
it's not inappropriate that what happened did happen, but it
was a true change in our assumptions as well. We were about
to go to Afghanistan, and we were looking around for all the
people it takes to do that. We had 89,000 people deployed in
the United States doing things that had nothing to do with the
DOD, but we happened to be the only ones around who could stand
around with guns in the airports and at border stations and so
on. Some of the things we were doing, like flying F-16s and
AWACS [Airborne Warning and Control System] and so on were a
little more specialized toward the DOD, and in fact we have
the role of protecting against enemy aircraft. We never thought
those aircraft would be flown by American Airlines.
So lots of changes have occurred, but I would say to you that
the kinds of problems that I have about integration of inform-
ation within the DOD are absolutely trivial compared to those
of whoever is going to have that job in the DHS. It's going
to be a very tough problem.
Oettinger: To underscore this, he's dealing with three services
and twenty-eight agencies, but it's a two-digit number, as opposed
to three-, four-, or five-digit numbers in terms of the
jurisdictions and responsibilities involved in that fragmented
homeland security environment.
Stenbit: One of the things that I worry about is that the DOD
has been asked to be the executive agent for a lot of national
things. One of them is the National Communications System [NCS],
which has to do with how communications are provided not for the
DOD, but for the nation in certain cases. For instance, there's
a system for senior government officials so that when all the
phones jam up on Mother's Day, for example, we have a system
that gets around that. It's called GETS [Government Emergency
Telecommunications Service] -- GETS gets around it. On 9/11,
the cell phones all blocked up. We've had, in the same NCS,
a proposal on the table for quite a while about how to unblock
cell phones for privileged users. That now exists in three
cities in the United States. The DOD is not going to do that
anymore, because NCS has transferred over to DHS. Since that
was done in the DOD as sort of an incremental additive to all
the stuff we were already doing, I think it's going to be very
hard for a long time for those functions to be done in DHS.
There are some interesting issues wrapped around what you
asked, not only on the intelligence side, but also on the
communications side.
Oettinger: It was created during the Kennedy or Johnson
administration.
Stenbit: But it was really reenergized when Judge [Harold]
Green broke up Ma Bell in 1982. All of a sudden, in a heart-
beat, we had to be able to negotiate with multiple vendors
in a legal way to create networks when a disaster strikes.
For instance, rebuilding the switches in Lower Manhattan was
entirely done by DOD folks through the NCS kinds of mechanisms,
although it was done for FEMA [Federal Emergency Management
Agency] and FEMA paid us. That could not have been done
without this infrastructure that's been around for a while.
Oettinger: This was augmented by the National Security
Telecommunications Advisory Committee [NSTAC]. It's an
interesting trend, because the NCS was sort of a paper tiger.
Then
------------------
- 11 -
the NSTAC brought in the chief executives of the increasingly
fragmented telecommunications industry, and over the last couple
of years people from the computer industry and the networks were
generally brought in as well. So the solutions to this problem
are not exactly right there, but they're working on them.
Stenbit: There are twelve outreaches from the DHS to industrial
groups for critical infrastructure protection. One of them is
telecommunications. That exists in the NSTAC. It's the only one
of the twelve that works. The other eleven-finance, food, energy,
chemicals…pick the twelve sectors that represent the basic
economy of the country-are pick-me-up teams that are trying to
figure out when to have a meeting and who's who.
Oettinger: Each of them has an Information Sharing and Analysis
Center that is supposed to coordinate among the various elements
of these industries within which the competitors won't give each
other the time of day. There are many unresolved issues.
Stenbit: I'm going to stop there. I could keep going, but it's
not my job and I'm happy I'm not over there.
Let's chat a bit about the issue of the budget and the
CONOPS and the n2 to n. I talked to you a bit about how twenty-
five years ago we were doing TRI-TAC [Tri-Service Tactical
Communications Program], which was trying to superimpose a
16-kilobit voice standard on everybody, and basically built
these big interoperability machines that cost an awful lot of
money. Nobody wanted them. They didn't work very well.
They were always late. The software didn't work. It was a
typical government program. Ultimately, though, had there been
a war in Europe, that's what would have been the glue that held
everything together.
Student: Didn't it get unglued in Desert Storm? When I started
with the Air Force, that stuff was still around.
Stenbit: I was going to say that we had started down the cell
phone world by Desert Storm, but we clearly hadn't facilitated
the whole place, so it was a mixture of the two.
Student: Was that the STU-III [secure telephone unit]?
Stenbit: No, STU-III is a wire-line system. It has nothing to
do with this. STU-IIs were 2.4-kilobit systems. That was a
different idea. Then we were in the world where AT&T had been
broken up and you could add stuff to the net, so now what
we needed to do was produce our own fancy handset, compared to
somebody else's handset. We had some solutions that we could
use, other than the ones that were prescribed before Judge
Green.
In any case, I was talking about 1975, and I would assert
that, in those days, like the Athenians when they were fighting
Sparta in 1000 B.C., we in military operations were forced to be
synchronous in space and time. Let me chat about that for a
second, because I think it's a big-picture principle. We were
forced to be constrained in time because everybody had to be in
the same place at the same time in order to have a fight. Think
of World War II movies, with the Navy battleships lined up off
the coast of the invasion beach going "One, two, three, NOW!"
- 12 -
WHAM! All the shells go. "Okay, one, two, three, NOW!" WHAM!
All the shells go. That is my definition of synchronization in
time and space.
We were synchronized in time because we were
in a switched telephone world that was run with erlangs(9) and
calculations and so on -- a decidedly smart push world. Anybody
who knew anything needed to know the total state of the universe
in order to figure out whom to tell -- or to push it to. When
the captain of the Pueblo was captured, if he had only known the
phone number of the Marines in Okinawa it would have been over in
half an hour. But what process would he ever have used to know
the phone number of the Marines in Okinawa? I would assert to you
that it was an impossible problem. It was made even more im-
possible by the fact that whatever radio he had would have been
different from whatever radio the Marines had.
That's what this n2 issue is. To have interoperability, if
there were n different devices you needed to have n2 solutions.
Each one had to have a solution. I guess it's really n times
n-1, so I'm off a bit. It's on the order of n2 . (I'm a Van
Gogh person, not a Norman Rockwell person. What's a factor of n
here and there?) We were forced to be synchronous in time,
because if somebody called you and you didn't answer, nothing
happened. We were forced to be synchronous in space, because
basically we didn't know whom to call, other than the ones who
were on the little plastic sheet we had in front of us. So while
you could be anywhere, you had to stay in the same place or else
nobody would be able to catch you. That was really what TRI-TAC
and some of these other systems were trying to fix -- to allow the
logical equivalent of telephones to move in the battlefield and
be useful. Certainly we never were able to connect aircraft,
other than in the most rudimentary of possible ways. We never
could correlate where they were with what we were saying, or
anything else. That was just impossible.
That has an implication, and it is that a guy named F.W.
Lanchester, a British engineer of the late nineteenth and early
twentieth-century wrote some stuff (you probably know this
better than I do, and I fully accept that if I'm off you'll help
me) that basically said that in a military engagement, whether
it's the Athenians and the Spartans or 1900 or the mid-1970s,
the defense has an advantage over the offense of three to one.10
It worked out over time. There were deviations from that. This
assumes they were more or less equally matched folks. If the
longbow comes out after the French horsemen, and there's a
technological leap, it doesn't work, but after it balances out
again it's a three-to-one kind of problem.
Why does that happen? It's easy. I don't know why it's
three, but I certainly know why it's an advantage: because the
defense gets to choose the place. Unless they're stupid, they
pick one that's good. They also have the smallest internal lines
of communication, so they can communicate better, by definition.
The bottom line, after you've wrapped all that together, is that
it's about three to one.
-------------------------
| |
|
Return to the opening page