August 23, 2013
The Berkeley Par Lab:
Progress in the Parallel Computing Landscape
David Patterson, Dennis Gannon, and Michael Wrinn
Editors
• The Network of Workstations (NOW) project, 1993-1998, developed NOW-I and NOW-II, which were clusters of workstations that proved valuable in applications ranging from encryption to sorting. The Inktomi search engine was first built on NOW II, which led to a search engine startup company. Inktomi Inc. in turn demonstrated to the fledgling Internet industry the value of clusters of a large number of low-cost computers versus fewer more expensive high-end servers, which Google and others followed.
Project alumni became faculty at Berkeley, Harvard, Illinois, Princeton, Rutgers, Stanford, Texas, and Wisconsin. Tom Anderson, Eric Brewer, David Culler, and I led the NOW project. 7
• NOW was controversial in that it argued for building scalable large-scale computers from standard networks and computers as compared to distributed, cache-coherent, non-uniform memory access time machines (CC-NUMA) as advocated by the Stanford DASH project and later embodied in the SGI Origin computers.
[7] A. Fox and D. Patterson. Engineering software as a service: An agile approach using cloud computing. Strawberry Canyon, 2013.
([ Google's twist (now known as Alphabet): Alphabet's Google put a twist on using off-the-shelf components, not only by applying scale-ability, but using statistical data, for examples, of expected-failure-rate of the harddrive components, and they took this further, because they have multiple massive data centers, they gather data on the actual failure-rate of different harddrive manufacturers, and they further optimize the purchasing decision on the particular manufacturer ─ because of scale and detail gathering, monitoring, and analysis of failure rate data ─ we could say large-scale users of commodity harddrive like Google have better data on harddrive reliability than the manufacturers themselves ─ the data knowledge of this type would also be applicable to data storage device like flash drive manufacturer. ])
([ Amazon Web Services (AWS) would also have operational statistical data on components failure rate ─ with moving parts being the most likely to fail like harddisk drive, then later to electronic parts like ... ])
([ I am not saying human are off-the-shelf components, because human are organism. However, if you were to imagine the collective humanity as one big giant ant colony, then in a sense, human beings are one of the many components, the raw material resource so-to-speak, that power The System. It is at this stage that I shall breached upon the idea of class or the tale of two cities. How to tell if you are part of the privileged class? See if you are treated like a human organism or you are treated like off-the-shelf commodity. ])
____________________________________
Defense Science Board Summer Study on Autonomy
June 2016
p.71 (pdf page 85)
For example, a study by Google found that disc drives with increased heat, noise, and read/write errors detected by self-monitoring, analysis, and reporting
technology (SMART) were 39 times more likely to fail, enabling active countermeasures, such as isolating failing sectors to prevent data loss or alerting maintainers to back up the disk. 48
48 E. Pinheiro, W. Weber, and L. Barroso, Failure Trends in a Large Disk Drive Population. Appears in the Proceedings of the 5th USENIX Conference on File and Storage Technologies (FAST’07) [February 2007].
source:
Report of the Defense Science Board
Summer Study on Autonomy
June 2016
[[ This report is a product of the Defense Science Board (DSB). ]]
[[ This report is unclassified and cleared for public release. ]]
filename: on autonomy (dsb June 2016).pdf
____________________________________
Russell L. Ackoff, Ackoff's best, 1999 [ ]
predicting the future and preparing for it (PtFaPfI)
pp.325-326
. . . .. : the type of model employed in OR (Operations Research) implies a particular paradigm of problem solving. It consists of two parts:
(1) predicting the future and
(2) preparing for it.
Clearly, the effectiveness of this approach depends critically on the accuracy with which the future can be predicted. It helps us a little, and may harm us much, to prepare perfectly for an imperfectly-predicted future.
Therefore, the paradigm of OR should be one that involves “designing a desirable future and inventing ways of bringing it about.” The future depends at least as much on what we and others do between now and then as it does on what has already happened. Therefore, we can affect it, and by collaboration with others--expanding the system to be controlled--we can increase our chances of “making it happen.” The wider the collaboration, the more closely we can approximate the future we have jointly designed. It is this perception by Fred Emery and Eric Trist that gave rise to their work in social ecology.
Prediction and preparation were the principal modalities of the Machine Age: design and invention are emerging as the principal modalities of the System Age. Prediction and preparation involve passive adaptation to an environment that is believed to be out of our control. Design and invention involve active control of a system's environment as well as the system itself.
The models currently employed by OR are evaluative in nature; they enable us to compare alternative decisions or decision rules that are “given.” In design and invention, however, the alternatives are “taken,” created. Creative solutions to problems are not ones obtained by selecting the best from among a well- or widely-recognized set of alternatives, but rather by finding or producting a new alternative. Such an alternative is frequently so superior to any of those previously perceived that formal evaluation is not required. If it is, however, then the evaluative models of OR may have a use. The challenge, therefore, is not so much to improve our methods of evaluation, but to improve our methods of design and invention.
The point of the views I have expressed up to this point is not that OR's concept of problem solving is useless, but that it should have been taken as a starting-point of OR's development, not as its end-point. To have taken it as the end-point was to have aborted OR's development and to have initiated its retreat from reality.
(Ackoff's best : his classic writings on management, Russell L. Ackoff., © 1999, pp.325-326.)
____________________________________
Mix it up
The "high-low mix" concept, born out of the group of Air Force reformers sometimes called the "Fighter Mafia"—Cols. John Boyd, Thomas Christie, Franklin “Chuck” Spinney, and Pierre Sprey—sought to fix the death spiral of fighter aircraft development cost and complexity by augmenting a few expensive, high-capability aircraft built for air superiority with a larger number of less-complicated and less-expensive aircraft that could handle basic air defense, strike missions, and close air support.
In the 1970s, the "high" was the F-15 Eagle, and the "low" was the multirole F-16 Fighting Falcon. The Navy followed suit with the F-14 Tomcat as its "high" and the F/A-18 Hornet as a less-expensive "low."
To fix the death spiral of fighter aircraft development cost
• a few expensive, high-capability aircraft built for air superiority
• "high" was the F-15 Eagle
• a larger number of less-complicated and less-expensive aircraft that could handle basic air defense, strike missions, and close air support.
• "low" was the multirole F-16 Fighting Falcon
source:
https://www.textise.net/showText.aspx?strURL=https://arstechnica.com/tech-policy/2021/04/sitrep-is-the-f-35-now-officially-a-failure/
____________________________________
• a larger number of less-complicated and less-expensive aircraft that could handle basic air defense, strike missions, and close air support.
• building scalable large-scale computers from standard networks and computers
• Network of Workstations (NOW) project, 1993-1998
• NOW-I and NOW-II, which were clusters of workstations
• the value of clusters of a large number of low-cost computers versus fewer more expensive high-end servers
____________________________________
• a few expensive, high-capability aircraft built for air superiority
• distributed, cache-coherent, non-uniform memory access time machines (CC-NUMA)
• Stanford DASH project
• SGI Origin computers
____________________________________
August 23, 2013
The Berkeley Par Lab:
Progress in the Parallel Computing Landscape
David Patterson, Dennis Gannon, and Michael Wrinn
Editors
• The Network of Workstations (NOW) project, 1993-1998, developed NOW-I and NOW-II, which were clusters of workstations that proved valuable in applications ranging from encryption to sorting. The Inktomi search engine was first built on NOW II, which led to a search engine startup company. Inktomi Inc. in turn demonstrated to the fledgling Internet industry the value of clusters of a large number of low-cost computers versus fewer more expensive high-end servers, which Google and others followed.
Project alumni became faculty at Berkeley, Harvard, Illinois, Princeton, Rutgers, Stanford, Texas, and Wisconsin. Tom Anderson, Eric Brewer, David Culler, and I led the NOW project. 7
• NOW was controversial in that it argued for building scalable large-scale computers from standard networks and computers as compared to distributed, cache-coherent, non-uniform memory access time machines (CC-NUMA) as advocated by the Stanford DASH project and later embodied in the SGI Origin computers.
[7] A. Fox and D. Patterson. Engineering software as a service: An agile approach using cloud computing. Strawberry Canyon, 2013.
____________________________________
Scale and scope
The dynamics of industrial capitalism
Alfred D. Chandler, Jr.
with assistance of Takashi Hikino
p.10
Obviously, too, the performance of an enterprise and its industry in one decade reflected investments made, personnel hired, technologies adopted, and markets obtained in the previous and earlier decades.
(Scale and scope : the dynamics of industrial capitalism / Alfred D. Chandler, Jr. with the assistance of Takashi Hikino., 1. big business--United States--history, 2. big business--Great Britain--history, 3. big business--Germany--history, 4. big business--Germany (west)--history, copyright © 1990, )
____________________________________
p.75
Empirically, sex, social class, and profession seem to be better predictors of someone's behavior than nationality.
(p.75, Blackswan, by Nassim Taleb (Nassim Nicholas Taleb), 2007)
____________________________________
No comments:
Post a Comment