f*******h
发帖数: 1269 | 1 What are the differences and connections? | g****t
发帖数: 39 | 2 Grid includes everything
Cluster for a local network
P2P mainly on data sharing in a decentralized manner, can be a implementation
mechanism for Grid
WS a more popular standard like RPC, CORBA etc
Grid and WS are converging
Semantic Grid/WS is the future? Welcome bricks
【在 f*******h 的大作中提到】
: What are the differences and connections?
| c******n
发帖数: 4965 | 3 Grid is just a buzz word, a hype,
in fact, even if there's anything new in "Grid",
it's just a unifying wrapper, giving u a
single interface to resources,
now everybody calls his stuff "grid",
it's funny because the guy Ian Foster got
so irritated by the attempt of other people
to proclaim their stuff "grid", that he wrote
a trash paper saying "only what and what may be called
grid ....", it's like merchants. this buzz word is
basically the result of properganda by this guy.
btw, WS and grid are
【在 g****t 的大作中提到】
: Grid includes everything
: Cluster for a local network
: P2P mainly on data sharing in a decentralized manner, can be a implementation
: mechanism for Grid
: WS a more popular standard like RPC, CORBA etc
: Grid and WS are converging
: Semantic Grid/WS is the future? Welcome bricks
| r*****s
发帖数: 985 | 4 "Grid is a type of parallel and distributed system that
enables the sharing, selection, and aggregation of
geographically distributed 'autonomous' resources
dynamically at runtime depending on their availability,
capability, performance, cost, and users'
quality-of-service requirements."
The resources on Grids are loosely coupled, which is the
same motivation of Web services. That's why Grid embraces
WS as its underlying mechanism.
WS can be either SOAP RPC/messaging styles, or uri-oriented
REST
【在 g****t 的大作中提到】
: Grid includes everything
: Cluster for a local network
: P2P mainly on data sharing in a decentralized manner, can be a implementation
: mechanism for Grid
: WS a more popular standard like RPC, CORBA etc
: Grid and WS are converging
: Semantic Grid/WS is the future? Welcome bricks
| r*****s
发帖数: 985 | 5 Grid, as an idea derived from power grids, itself is good.
We should give Ian some credits as he and Carl did a wonderful
marketing job at a right time. Why isn't legion or condor but
globus as the first important grid product? There must be
some reasons in it ;)
BTW, currently,
Grid Service == WS + OGSA
【在 c******n 的大作中提到】
: Grid is just a buzz word, a hype,
: in fact, even if there's anything new in "Grid",
: it's just a unifying wrapper, giving u a
: single interface to resources,
: now everybody calls his stuff "grid",
: it's funny because the guy Ian Foster got
: so irritated by the attempt of other people
: to proclaim their stuff "grid", that he wrote
: a trash paper saying "only what and what may be called
: grid ....", it's like merchants. this buzz word is
| c******n
发帖数: 4965 | 6 things like globus are evaluated by their acceptance rate
not by their technical creativeness or techincal contribution,
globus is really a simple concept, and its capability is implemented
at least in condor, among others, there's nothing new.
public acceptance has large part to do with Marketing,
like the acceptance of DOS over, say, Unix. and like all protocols,
is the result of political bargainings between the parties involved
if there is "some reasons" for the acceptance of Globus, then it
【在 r*****s 的大作中提到】
: Grid, as an idea derived from power grids, itself is good.
: We should give Ian some credits as he and Carl did a wonderful
: marketing job at a right time. Why isn't legion or condor but
: globus as the first important grid product? There must be
: some reasons in it ;)
: BTW, currently,
: Grid Service == WS + OGSA
| g****t
发帖数: 39 | 7 Some more reasons:
1. globus is simpler (like DOS is simpler). Now it is reduced to be a
SSH-alike or similar stuff. It leaves space for others to survive and join,
although GGF is far less effective than IETF.
2. globus is so-called open. Because it is simple and not a working
implementation, it has to be open to enable other functions, while condor is a
fully working solution.
3. marketing and resound the hype for open source, open standard. In addition,
it survives from the dillema between .N
【在 c******n 的大作中提到】
: things like globus are evaluated by their acceptance rate
: not by their technical creativeness or techincal contribution,
: globus is really a simple concept, and its capability is implemented
: at least in condor, among others, there's nothing new.
: public acceptance has large part to do with Marketing,
: like the acceptance of DOS over, say, Unix. and like all protocols,
: is the result of political bargainings between the parties involved
: if there is "some reasons" for the acceptance of Globus, then it
| g****t
发帖数: 39 | 8 Anyone has good pointers to some really big guys with real and detailed
innovations, not just engineering a framework? Particularly on resource
management and fault-tolerance etc.
Condor's match-maker seems quite impressive. Any other good ones? Foster's
papers are good at engineering or integrating some frameworks and make them
well-known or well-accepted. But I feel very hard to learn from those papers.
In my opinion, you have to be already a big NIU to be able to publish those
papers.
Any tho | r*****s
发帖数: 985 | 9 not sure which version of Globus are you talking about, GT2?
GT3 and 4 are totally Web services.
GT3 = Web services + OGSI(OGSA)
GT4 = Web services + WSRF(OGSA)
BTW, Globus is not grid. it is a set of toolkits. There are many different
implementations, as long as it follows OGSA.
Sometimes, if you can coordinate everybody to use one standard, you are
the hero. Not every best solution is adopted. Fights between big companies
really hurt technologies.
【在 g****t 的大作中提到】
: Some more reasons:
: 1. globus is simpler (like DOS is simpler). Now it is reduced to be a
: SSH-alike or similar stuff. It leaves space for others to survive and join,
: although GGF is far less effective than IETF.
: 2. globus is so-called open. Because it is simple and not a working
: implementation, it has to be open to enable other functions, while condor is a
: fully working solution.
: 3. marketing and resound the hype for open source, open standard. In addition,
: it survives from the dillema between .N
| r*****s
发帖数: 985 | 10 take a look at the major conferences on Grid like Grid Workshop on SC,
CCGrid, HPDC ...
【在 g****t 的大作中提到】
: Anyone has good pointers to some really big guys with real and detailed
: innovations, not just engineering a framework? Particularly on resource
: management and fault-tolerance etc.
: Condor's match-maker seems quite impressive. Any other good ones? Foster's
: papers are good at engineering or integrating some frameworks and make them
: well-known or well-accepted. But I feel very hard to learn from those papers.
: In my opinion, you have to be already a big NIU to be able to publish those
: papers.
: Any tho
| | | c******n
发帖数: 4965 | 11 这个所谓Grid 概念反映了Ian Foster 象商人一样的作法,
本来distributed computing ( not counting theoretical research )的目的
就是把各种resources 连起来以能easy access, 或者说这行业的人
一直都在做这个, 譬如condor 做的就是,
打个比方, 大家都会吃苹果,也从小到大都在吃苹果,
现在有张三突然说 "我提出一个吃苹果的framework, 这是一个open
standard, 你要先拿起来,然后看准,然后咬..... "
你不觉得很可笑么?
再说这个open 的概念吧, 任何东西没有写完都可以说是open,
你照样可以写些小程序在condor 的framework下面work,
也可以叫做extentions to that open framework ......
再看globus 2 到3的彻头彻尾的变化, 只有concept 上的继承,
或者说open 的只是concept/architecture,
【在 r*****s 的大作中提到】
: not sure which version of Globus are you talking about, GT2?
: GT3 and 4 are totally Web services.
: GT3 = Web services + OGSI(OGSA)
: GT4 = Web services + WSRF(OGSA)
: BTW, Globus is not grid. it is a set of toolkits. There are many different
: implementations, as long as it follows OGSA.
: Sometimes, if you can coordinate everybody to use one standard, you are
: the hero. Not every best solution is adopted. Fights between big companies
: really hurt technologies.
| j******e
发帖数: 232 | 12 你们都很恨Ian Foster吗
讲讲rumors吧
【在 c******n 的大作中提到】
: 这个所谓Grid 概念反映了Ian Foster 象商人一样的作法,
: 本来distributed computing ( not counting theoretical research )的目的
: 就是把各种resources 连起来以能easy access, 或者说这行业的人
: 一直都在做这个, 譬如condor 做的就是,
: 打个比方, 大家都会吃苹果,也从小到大都在吃苹果,
: 现在有张三突然说 "我提出一个吃苹果的framework, 这是一个open
: standard, 你要先拿起来,然后看准,然后咬..... "
: 你不觉得很可笑么?
: 再说这个open 的概念吧, 任何东西没有写完都可以说是open,
: 你照样可以写些小程序在condor 的framework下面work,
| g****t
发帖数: 39 | 13 Why hate? We like him :). He opened a broader way for us to have a possibly
better career.
Without Grid, everybody in distributed computing will work on either J2EE or
.Net that are undergraduates' world :).
different
companies
【在 j******e 的大作中提到】
: 你们都很恨Ian Foster吗
: 讲讲rumors吧
| j******e
发帖数: 232 | 14 可是感觉他的工作就是定标准定框架
既没有理论贡献也没有什么特别好的直接应用
似乎属于那种谁来干都一样只不过他先干的工作
我瞎说啊
【在 g****t 的大作中提到】
: Why hate? We like him :). He opened a broader way for us to have a possibly
: better career.
: Without Grid, everybody in distributed computing will work on either J2EE or
: .Net that are undergraduates' world :).
:
: different
: companies
| g****t
发帖数: 39 | 15 Thank you. I noticed these conferences also.
Any good one on analytical models? Typical papers on Grid are as follows:
1. a framework, preliminary implementation, some application experiments, show
several tables and figures, then give some intuitive thinking
2. an algo, with complex and fancy flowchart, some experiments, a little bit
on why and how
Any papers show the upper bound on the performance limit, speedup/efficiency,
analytical models, stochastic analysis etc. I noticed one stochastic p
【在 r*****s 的大作中提到】
: take a look at the major conferences on Grid like Grid Workshop on SC,
: CCGrid, HPDC ...
| j******e
发帖数: 232 | 16 我还以为你们谁是U Chicago的呢
看来都不是
没有rumor不好玩
【在 g****t 的大作中提到】
: Thank you. I noticed these conferences also.
: Any good one on analytical models? Typical papers on Grid are as follows:
: 1. a framework, preliminary implementation, some application experiments, show
: several tables and figures, then give some intuitive thinking
: 2. an algo, with complex and fancy flowchart, some experiments, a little bit
: on why and how
: Any papers show the upper bound on the performance limit, speedup/efficiency,
: analytical models, stochastic analysis etc. I noticed one stochastic p
| g****t
发帖数: 39 | 17 Without him, few guys have the energy to push the Grid concept. He has been
involved in many grande projects on high-energy physics, bioinformatics etc.
An emerging industry or innovation needs one visioner or agenda-setter or a
group.
Regarding the innovation, the high-profiled standardization process is a big
innovation in Grid. It could result in some breakthrough in the long-term with
millions of researchers and billions $, who knows.
Forget these general and big vision, what can we do to co
【在 j******e 的大作中提到】
: 可是感觉他的工作就是定标准定框架
: 既没有理论贡献也没有什么特别好的直接应用
: 似乎属于那种谁来干都一样只不过他先干的工作
: 我瞎说啊
| c******n
发帖数: 4965 | 18 papers fall into these areas
(1) assume an abstract model, describe it as
some theoretical problem, formalize it.
prove it's NPC.... give approximation alg, give
heuristics
show some simulations results
(2) small implementation project,
adding a particular component into the framework,
improve upon existing ones, such as resource discovery ....
(3) a particular application, for example, in medicine, in oceanography ....
(4) a lot of work on resource dis
【在 g****t 的大作中提到】
: Thank you. I noticed these conferences also.
: Any good one on analytical models? Typical papers on Grid are as follows:
: 1. a framework, preliminary implementation, some application experiments, show
: several tables and figures, then give some intuitive thinking
: 2. an algo, with complex and fancy flowchart, some experiments, a little bit
: on why and how
: Any papers show the upper bound on the performance limit, speedup/efficiency,
: analytical models, stochastic analysis etc. I noticed one stochastic p
| g****t
发帖数: 39 | 19 Grid is not only for fun :). It can be though, such as mass game platform with
millions of players.
Don't pay too much attention to rumors. He did great contribution already. We
have to ask ourselves what we can do?
show
bit
speedup/efficiency,
paper. I
【在 j******e 的大作中提到】
: 我还以为你们谁是U Chicago的呢
: 看来都不是
: 没有rumor不好玩
| c******n
发帖数: 4965 | 20 ft, if you claim online multiplayer games to be "grid"
I fool eh you
sony 还真吹过它的PS3要grid-enabled 如何如何, 现在也不吹了
【在 g****t 的大作中提到】
: Grid is not only for fun :). It can be though, such as mass game platform with
: millions of players.
: Don't pay too much attention to rumors. He did great contribution already. We
: have to ask ourselves what we can do?
:
: show
: bit
: speedup/efficiency,
: paper. I
| | | j******e
发帖数: 232 | 21 读书最怕固执
we can build anything on grid
online game is definitely one possible application
【在 c******n 的大作中提到】
: ft, if you claim online multiplayer games to be "grid"
: I fool eh you
: sony 还真吹过它的PS3要grid-enabled 如何如何, 现在也不吹了
| g****t
发帖数: 39 | 22 why not? look at the following
http://www.cis.upenn.edu/~hhl/games/
With Grid (or any else) standard, all game vendors can join the game sever
pools.
Grid is nothing but resource virtualization and management in a standard way.
with
We
【在 c******n 的大作中提到】
: ft, if you claim online multiplayer games to be "grid"
: I fool eh you
: sony 还真吹过它的PS3要grid-enabled 如何如何, 现在也不吹了
| r*****s
发帖数: 985 | 23 no rumors ... the old news is that he and carl created
a company called UNIVA.
he himself is a very smart guy. a model for us,
with such a successful career in such an age,
a happy family -- a beautiful wife, one cute
daughter and one son...
BTW, before 1997, it's called Metacomputing ...
【在 g****t 的大作中提到】
: Why hate? We like him :). He opened a broader way for us to have a possibly
: better career.
: Without Grid, everybody in distributed computing will work on either J2EE or
: .Net that are undergraduates' world :).
:
: different
: companies
| r*****s
发帖数: 985 | 24 Well, his contribution feeds so many professors and students all over the
america/world including me, with funds, degrees, tenures ... :D
I would suggest not to think too much on the world "Grid". All the general
knowledge/theories/techniques apply to "Grid". If you really want to get
one step closer, that's Web services. There are so open issues in WS, but
industry does it better than academia ...
【在 g****t 的大作中提到】
: Without him, few guys have the energy to push the Grid concept. He has been
: involved in many grande projects on high-energy physics, bioinformatics etc.
: An emerging industry or innovation needs one visioner or agenda-setter or a
: group.
: Regarding the innovation, the high-profiled standardization process is a big
: innovation in Grid. It could result in some breakthrough in the long-term with
: millions of researchers and billions $, who knows.
: Forget these general and big vision, what can we do to co
| s*****o
发帖数: 1540 | 25 Introducing a book on P2P, WS, Grid, and even better, containing all of them:
haha..
From P2P to Web Services and Grids: Peers in a Client/Server World
by Ian J. Taylor
http://www.amazon.com/exec/obidos/tg/detail/-/1852338695/qid=1096920665/sr=1-1
/ref=sr_1_1/104-1319831-5335111?v=glance&s=books
been
etc.
a
big
with
something
and
【在 r*****s 的大作中提到】
: Well, his contribution feeds so many professors and students all over the
: america/world including me, with funds, degrees, tenures ... :D
: I would suggest not to think too much on the world "Grid". All the general
: knowledge/theories/techniques apply to "Grid". If you really want to get
: one step closer, that's Web services. There are so open issues in WS, but
: industry does it better than academia ...
| g****t
发帖数: 39 | 26 Thank you. I hit this book very recently but have not a chance to read it yet.
It sounds too broad though. Any insight on it? Any PDF ebook :)?
them:
http://www.amazon.com/exec/obidos/tg/detail/-/1852338695/qid=1096920665/sr=1-1
or
long-term
(small
【在 s*****o 的大作中提到】
: Introducing a book on P2P, WS, Grid, and even better, containing all of them:
: haha..
: From P2P to Web Services and Grids: Peers in a Client/Server World
: by Ian J. Taylor
: http://www.amazon.com/exec/obidos/tg/detail/-/1852338695/qid=1096920665/sr=1-1
: /ref=sr_1_1/104-1319831-5335111?v=glance&s=books
:
: been
: etc.
: a
| f*******h
发帖数: 1269 | 27 I am looking a good book on Grid too.
Anybody can recommend some classical papers to read? Thanks!
yet.
http://www.amazon.com/exec/obidos/tg/detail/-/1852338695/qid=1096920665/sr=1-1
the
general
but
bioinformatics
a
【在 g****t 的大作中提到】
: Thank you. I hit this book very recently but have not a chance to read it yet.
: It sounds too broad though. Any insight on it? Any PDF ebook :)?
:
: them:
: http://www.amazon.com/exec/obidos/tg/detail/-/1852338695/qid=1096920665/sr=1-1
: or
: long-term
: (small
| r*****s
发帖数: 985 | 28 globus.org 有一堆paper.
now we don't talk much about Grid, but Web services instead ...
GT3-->GT4 really drives a lot of ppl nut, especially those brits :D
【在 f*******h 的大作中提到】
: I am looking a good book on Grid too.
: Anybody can recommend some classical papers to read? Thanks!
:
: yet.
: http://www.amazon.com/exec/obidos/tg/detail/-/1852338695/qid=1096920665/sr=1-1
: the
: general
: but
: bioinformatics
: a
| s*****o
发帖数: 1540 | 29
A classic book:
The Grid 2: Blueprint for a New Computing Infrastructure
by Ian Foster, Carl Kesselman
http://www.amazon.com/exec/obidos/ASIN/1558609334/qid=1106776502/sr=2-1/ref=pd
_ka_b_2_1/103-4464603-2996623
Paper? hehe... Google ~
http://www.amazon.com/exec/obidos/tg/detail/-/1852338695/qid=1096920665/sr=1-1
get
has
agenda-setter
is
【在 f*******h 的大作中提到】
: I am looking a good book on Grid too.
: Anybody can recommend some classical papers to read? Thanks!
:
: yet.
: http://www.amazon.com/exec/obidos/tg/detail/-/1852338695/qid=1096920665/sr=1-1
: the
: general
: but
: bioinformatics
: a
| w********g
发帖数: 121 | 30 nice discussion, why don't u guys go on and on?
【在 s*****o 的大作中提到】
:
: A classic book:
: The Grid 2: Blueprint for a New Computing Infrastructure
: by Ian Foster, Carl Kesselman
: http://www.amazon.com/exec/obidos/ASIN/1558609334/qid=1106776502/sr=2-1/ref=pd
: _ka_b_2_1/103-4464603-2996623
: Paper? hehe... Google ~
: http://www.amazon.com/exec/obidos/tg/detail/-/1852338695/qid=1096920665/sr=1-1
: get
: has
|
|
