E-notebook middleware for accountability and reputation based trust in distributed data sha

E-notebook middleware for accountability and reputation based trust in distributed data sha
E-notebook middleware for accountability and reputation based trust in distributed data sha

E-notebook Middleware for Accountability and Reputation Based Trust in Distributed Data

Sharing Communities

Paul Ruth ,Dongyan Xu ?,Bharat Bhargava ?,and Fred Regnier?

Purdue University,West Lafayette,https://www.360docs.net/doc/771694625.html,A,

{ruth,dxu,bb}@https://www.360docs.net/doc/771694625.html,,fregnier@https://www.360docs.net/doc/771694625.html,

Abstract.This paper presents the design of a new middleware which

provides support for trust and accountability in distributed data sharing

communities.One application is in the context of scienti?c collabora-

tions.Multiple researchers share individually collected data,who in turn

create new data sets by performing transformations on existing shared

data sets.In data sharing communities building trust for the data ob-

tained from others is crucial.However,the?eld of data provenance does

not consider malicious or untrustworthy users.By adding accountability

to the provenance of each data set,this middlware ensures data integrity

insofar as any errors can be identi?ed and corrected.The user is further

protected from faulty data by a trust view created from past experiences

and second-hand recommendations.A trust view is based on real world

social interactions and re?ects each user’s own experiences within the

community.By identifying the providers of faulty data and removing

them from a trust view,the integrity of all data is enhanced

1Introduction

In scienti?c research,scientists rely on experimental data to demonstrate their ?ndings.The accuracy of the data is critical not only for the validity of the research results but also for the reputation of the scientist.Currently,a scientist’s professional reputation is determined by peer review of papers submitted to conferences and journals for publication.Frequently,results obtained are based on complete data that does not accompany the paper.It is assumed that the integrity of the data has been maintained throughout.

To complicate matters even more,the recent growth in processing power and storage capacity along with the ease of communication through the Internet, has allowed scientists to create and process very large data sets based on locally derived data as well as data obtained from other scientists.Although a large data set can provide better results because of larger and more diverse sampling,

in order to be con?dent with the results,the origin of all data in the set must be known.In most scienti?c communities there is no standardized method for collecting and sharing data,which makes it di?cult to achieve global data con-sistency,validity,and credibility.More speci?cally heterogeneity between labs may lay in the following:

–The condition and calibration of experimental instruments in di?erent labs, and the condition and con?guration of lab environments.

–The context of di?erent experiments,such as the time,location,temperature of the experiments,and in the case of medical or social experiments,the age and ethnic group of human subjects.

–The protocol(and the strictness of its enforcement)of data generation,trans-formation,and derivation.For example,di?erent labs may use di?erent sam-pling rates,precision,and number of repetitions.

–The capacity,version,and con?guration of computing platforms(both soft-ware and hardware)in di?erent labs.

–Non-uniform data formats adopted by di?erent labs,due to their formatting conventions and di?erences in software/hardware/instruments.

There is a need for a distributed environment that allows researchers to collaborate by sharing data while maintaining the complete history and source of all data sets.This by necessity would include those smaller sets which constitute the greater accumulation of data and the transformations from which they were combined.The?eld of data provenance is evolving out of this concern[2,4,6–8,10–12,14,16,17,20,25,27].Data provenance is the description of the origins of a piece of data and the process by which it arrived in a database[7].Data provenance is often used to validate data or re-execute a derivation with di?erent input parameters.Currently the?eld of data provenance is working on how to annotate large ad-hoc data sets in order to identify and correct erroneous data or rederive data sets based on new input.However,the existence of malicious and incompetent users has not been considered.To date,data provenance projects have considered all participants to be trustworthy and meta-data to be correct.

We have determined that data provenance schemes can also be used to store information regarding the validity of data sets.Similar to how the scienti?c community performs peer reviews on scienti?c research,shared data sets can be subjected to peer review before they are widely https://www.360docs.net/doc/771694625.html,ers of the shared data set will be able to assess the set’s integrity through a similar analytic process as that employed in the peer review process and malicious or incompetent users will be exposed.

In most?elds of science,instruments for collecting data and the algorithms to operate on data are constantly advancing.Ideally,any system which expedites the communal sharing of data should record all of the context information related to the data’s collection and transformation.By using our system,an individual scientist may investigate the history of a particular data set to determine if s/he disagrees with any collection techniques or transformation algorithms used to construct it.The scientist could then explore whether users with a previously

determined low reputation collected or derived any part of the data.Thus,by allowing the examiner to asses the product as a sum of its parts,s/he can produce a thorough peer review of the data.

It is important to investigate the source of collaborative data.Errors made at very low levels may never be seen once the data is integrated and replicated multiple times.One might consider the a?ect a miscon?gured instrument may have on data obtained in any given data set.Unless the con?guration/calibration of each instrument used to collect the data is recorded,it may never be possible to identify the problem at a later stage.

A more speci?c example is found in bioinfomatics.Here the functional anno-tation of proteins by genome sequencing projects is often inferred from similar, previously annotated proteins.The source of the annotation is often not recorded so annotation errors can propagate throughout much of the data base[5,9,13, 15,21].In extreme cases,data may be faked intentionally.In2002,an external committee concluded that former Bell Labs researcher Hendrik Sch¨o n.manip-ulated and misrepresented data in16papers involving20co-authors[3].These co-authors and an unknown number of scientists who have used parts of the16 falsi?ed papers all blindly trusted the integrity of Sch¨o n’s data.It has been sug-gested,in the wake of this incident that the research community adopt a data auditing and validation system which can help verify the integrity of data and results independently.

We have designed a system that records the history of a data set similar to other data provenance systems which use a directed acyclic graph(DAG). However,our system establishes a cryptographic signature for each data set and its history.A user will then be accountable for the validity of each signed data set.If a data set contains material contributed by many other sources,the identity of those sources will be included in the history of the larger set.In this way,not only can the original faulty data set be found,but the researcher who made the mistake can be held accountable.

Once a system of accountability is in place,a trust management system based on real world notions of trust and reputation can be implemented.This will go a long way towards increasing the probability that an individual data set is valid and increase the integrity of the data in the entire https://www.360docs.net/doc/771694625.html,ers will interact with each other and record their experiences.Each user will individually evaluate the probable integrity of each piece of data based on the unforgeable and irrefutable information contained in the signed histories,his or her personal experiences,and the recommendations of others.

The remainder of this paper is organized as follows:section2gives a brief summary of current related work,section3provides an overview of our project goals,section4shows the basic architecture of the system,section5provides the data history data structure,section6gives a detailed description of how trust views are determined and implemented,and the?nal sections talk about future work and our conclusions.

2Related Work

Recently there has been increased activity in the area of data provenance.Re-search is focused on providing the correct annotations for recording the history of data.What follows is a very brief description of several data provenance projects.

The Chimera project[2,10,11]is is a system developed as part of the Grid Physics Network(GriPhyN)project.Chimera provides for ad-hoc sharing and creation of distributed data sets while recording the history of each data set.The purpose of Chimera is to play the role of a make?le for large systems comprised of many data sets distributed across a network.This distributed make?le allows for the recreation of large data sets when a smaller sub-data set is changed. However,Chimera neither provides accountability for the shared data,nor helps users determine which data sets are most likely to consist of valid information.

Earth System Science Workbench(ESSW)[4,12]is for data centers studying earth sciences.The goal of the project is to allow participating data centers to search and obtain data while publishing their own data.This project does not consider malicious or incompetent users in the system.

The my Grid project[16,25,27]has more capabilities.my Grid is a complete e-Science system in which not only data will be shared but all electronic resources including:instruments,sensors,data,and computational methods.In essence, my Grid provides for an experiment to be done completely in silico.However the

data in the system is assumed to be correct and of high integrity.

ESP2Net[20]is developing a Scienti?c Experiment Markup Language(SEML). SEML is based on XML and is a language which requires data provenance in-formation be stored with all data.SEML is aimed at scienti?c data sharing.

The PENN Database Research Group led by Peter Bunemen[7]has done signi?cant work at the lower levels of data provenance.Their work is focused on how to record data provenance within a database and does not consider the peer-to-peer relationships formed by the various data providers.

Audun J?sang[18,19,24]has concentrated his research on the theoretical side of trust.Most of his work in the logic of trust relationships.More recently he has studied the trust relationship between agents in e-commerce.

Alfarez Abdul-Rahman[1]proposed a model of trust that mimics the real world trust each of us exhibits everyday when dealing with other people.His trust model allows for each participant to form their own opinion of other peers based on his or her own experiences with the system.Each user will independently form this opinion and the opinion with change as more experiences are created. We use Abdula-Rahman’s trust model in our system.

RAID lab[26]has developed an approach to establish the trust of a prin-cipal,Alice,based on her history(i.e.a sequence of trust establishment events that involved Alice).They assume that Alice obtains a rating for each event that characterizes her behavior.Their approach is context-sensitive in that it considers the ratings and attributes associated with trust establishment events such as risk and event sequence patterns.

3Overview

The goal of our research is to design and prototype an accountable,trust-aware, and data-centric e-notebook middleware.This e-notebook middleware is dis-tributed,running on machines in individual research labs and possibly on larger servers(for example a campus-wide e-notebook that could be created at a univer-sity or large company).The e-notebook will record(1)the context in which raw data is generated(by communicating with on-board software)and(2)the history of curated data including data transformation,derivation,and validation.The individual,through his or her e-notebook,will digitally sign and be accountable for the result of every process performed.Based on the information recorded and experiences with others participating in the network,the distributed e-notebook will establish and maintain trust views for scientists sharing scienti?c data.We contend that these trust views and accountability for each data item will provide a measure of con?dence in the shared data similar to the trust gained by the peer review process.

The e-notebook will change the way scienti?c data is compared and corre-lated.With the proposed e-notebook,a user will not only judge the value of a data set,the context in which the data was collected and the history(organized as a directed acyclic graph recording the steps of data collection and transforma-tion from the very beginning)of how the data came to be can be used,improving the trustworthiness of scienti?c discoveries based on such comparison.

e?notebook

Fig.1.High level view of the https://www.360docs.net/doc/771694625.html,ers will participate in the community through an e-notebook.Each e-notebook will store some amount of data and participate in the querying for and sharing of data.In addition,each e-notebook will digitally sign and claim responsibility for data sets which it creates.

4Architecture

The middleware architecture(?gure1)of the proposed system will be highly distributed and?exible.The key element of the system is the e-notebook.Each user will create his or her own notebook and through it collaborate with other users by querying for and retrieving data published on their e-notebooks.The access is also provided to instruments for collecting raw data.The e-notebook will do more than simply collect raw data from the instruments.It will also collect all contextual data(instrument settings,temperature,time,researcher’s name,etc.)that the researcher might not think are important.Similar to other data provenance research projects the desired way to accomplish this is for the e-notebook to be connected directly to the instrument’s on-board software.It will also be possible for a researcher to input data manually.It should be noted, however,that human error and the common desire to exclude seemingly irrele-vant data demonstrates the bene?t of automating this process.The e-notebook will also record all applications of transformations on a data set.

In addition to e-notebooks which belong to individual scientist,there may be e-notebooks that reside on servers for the purpose of sharing large amounts of data.An e-notebook of this type will be identical to a regular one and provide a sharing and storage facility for a group of users.Ideal sites for a server e-notebook may include universities and large companies.The only di?erences between a user e-notebook and a server e-notebook will be the size and the way that it is used.Server e-notebooks will have a larger storage capacity and higher bandwidth capabilities.A server e-notebook’s intent is to provide a large repository for storing data that regular users might not want to store locally. The server e-notebook will query for and download any and all data which is to be shared.It may be desirable for the owner of a server e-notebook to allow other users to upload data to the server themselves.

5Data History and Evidence

When data is collected,transformed,and combined in a distributed ad-hoc man-ner by di?erent people with di?erent agendas,the temporal history of the data is often lost.Data provenance is the recording of meta-data which describes the history of a data set.Our design of the data provenance system not only records the history of the data,but extends the current systems to include unforgeable and irrefutable evidence of what happened to the data and who performed those actions.

We use a data provenance scheme,similar to current data provenance sys-tems[2,4,6–8,10–12,14,16,17,20,25,27]in which a directed acyclic graph(DAG) is used to describe a data set’s history(?gure2).In our design,a data set’s DAG is a digitally signed claim of the history of a data set made by the user of the e-notebook from which it was created.Each node in the DAG contains a sin-gle data set and information describing how it was created.Some data sets are collected directly from instruments while others are created by performing trans-formations on one or more existing data sets.For each data set created through

transformations there will be a directed edge from the data set’s node to each node used as input to the transformation.In?gure2,data sets1-3were collected

transformations.

directly from instruments while data sets4-6were the results of

When a user collects or derives a new data set,s/he creates a new node with directed edges to a copy of each node used to create the new one.The entire DAG is digitally signed by its creator.It is worth mentioning that within the signed DAG each sub-DAG remains digitally signed by and accountable to its original creator.

The purpose of digitally signing the DAG is to establish accountability.When a user signs a DAG,s/he claims creative responsibility and credit for the data. All sub-DAGs will remain signed by and accountable to their creators.The DAG is then published through the e-notebook for download and use by other users.

When a user downloads a data set,that user may wish to investigate its history by searching the DAG.In this manner,s/he can know all transformations which were applied,all users who were involved,and the context in which the data was collected.It can be known if any of the transformations,users,or contexts are inadequate for the intended use of the data set and if necessary,the material may be avoided.

In some cases,downloaded data sets may contain errors(intentional or other-wise).If an error is found,the digitally signed DAG is unforgeable and irrefutable evidence of what happened and who is responsible.At the very least,the evi-dence of errors in data sets should be used to adjust the reputation of the careless or malicious user,while at the most the evidence can be made public(possible in a court case)to prove malice.Although,the intent of the system is to increase the integrity of all data in the system by discouraging inappropriate use of the system,the evidence of carelessness and malice must be strong enough to present in court for this to be e?ective.

Figure2shows an example DAG for data set6,which was created and signed by user A.With his signature A is claiming that he created data set6 using transformation3and input data sets4and5.Both data sets4and5are in turn signed by their creators.In this case data set4happens to have been created and signed by A who performed the transformation to create data set6. The other input data set,5was created and signed by B.Because data set5is signed by B,A makes no claims to its validity.A only claims that he agreed to the use of data set5.If data set5,or any data which went into it,is discovered to be faulty,user A should disband the use of that data set and the creator of the?rst faulty data set is held accountable.

If any user were to obtain data set2and3along with transformation2, s/he can validate user A’s claim by recreating data set5.If it is not possible to recreate data set5by applying transformation2to data sets2and3,user A did not create data set5this way and incorrectly made the claim that s/he did. Once user A digitally signs data set5’s DAG and releases it to the community, user A can never assert s/he did not make this claim.If it can be shown that data set5was not created in the way user A claimed it was,the signed DAG is evidence that A released incorrect data to the community.Evidence of malice cannot be shown with the DAG and must be determined in some other way.

6Trust Views

One novel technique used by our system is the formation of trust views resulting from the reputation of an e-notebook https://www.360docs.net/doc/771694625.html,ing previous,?rst-hand experience and second-hand recommendations each user will decide how to trust other e-notebook users.As in real world situations involving trust,there is no universal value assigned to the integrity of each user.No person necessarily judges in-tegrity in the same way as someone else.Each user may have his own algorithm for determining the integrity of others.We propose that using the signed history DAGs described in section5users have enough information to make value judg-ments.This will increase the probability of an individual obtaining valid data and raise the integrity of all the data in the system.

6.1Trust Judgments

In order to create a trust view,each user must make judgments of how much and what kind of trust to assign other users.E-notebook users can make trust judg-

ments in any way they wish.At?rst,users might rely on o?-line relationships. However,as experiences with the community increases,it becomes possible to use accountability information obtained from the signed data histories to make judgments about other’s?ndings.There are endless possibilities in which to use signed histories to make trust judgments.Describing them all would be impos-sible.Listed below are a few properties which might lead to an increase in the level of trust assigned to a user:

–Consistently producing mistake free data sets.

–Quickly modifying data when mistakes are found in lower level data sets.–Recommending users who provide quality data sets.

Alternatively,the next list of properties might lead to a reduction of a user’s trust:

–Creating and signing a data set which is known to be intentionally fraudulent.–Consistently making unintentional mistakes in the creation of new data sets.–Using data which are known to be faulty in the creation of new data sets.–Recommending users who provide faulty data sets.

In addition to personal experiences,trust judgments can be made using sec-ond hand recommendations.Building trust in recommendations can initially be done by accepting the positive assessments of other users who are known out-side of the system.Once a base of trust has been established,one may trust the recommendation of users who are unknown outside the system.

Abdul-Rahman describes one social model for supporting trust in virtual communities[1].In this research,agents trust each other by ranking all?rst-hand experiences into discrete categories(for example:very good,good,bad, very bad).If only?rst-hand experiences were considered,when deciding on the trust to award another agent the trust category with the most experiences in it is used.However,Abdul-Rahman provides for trusting through recommen-dations as well.Recommendations are made by sharing assessments based on ?rst hand-experiences.However,an agent cannot use recommended experiences in the same way as?rst-hand experiences.The technique used is to calculate the semantic di?erence between recommendations received and?rst-hand expe-riences using those recommendations.Future recommendations can be modi?ed by the semantic di?erence seen in the past to more accurately suggest amounts of trust to award.In other words,for each user who makes recommendations,the receiving users will calculate the typical di?erence between the recommendation and personally observed outcome.The typical di?erence can then be applied to adjust future recommendation from that user.

We have designed a similar model of social trust for users to determine the probability that a given data set is valid.In our system,agents are users and the categories are very trustworthy,trustworthy,untrustworthy,and very untrust-worthy.It should be noted that any?nite number of categories will work and we chose four categories to mirror Abdul-Rahman’s work.Each user will record all?rst hand experiences and determine which category each experience should

belong to.At any given time,the trust level determined by?rst hand experi-ence is the level associated with the category containing the most experiences. For example,if user A has4very trustworthy experiences and5trustworthy experiences with user B,then A applies the category trustworthy to B.

Recommendations are made by incorporating the experiences of others into one’s rating.Each user has his or her own experiences and techniques for cate-gorizing the experiences.For this reason,another user’s recommendation must be adjusted to approximately?t his or her categorizations.To do this the past recommendations and the user’s resulting experiences are used to?nd the seman-tic di?erence between the recommendations and his or her experiences.This is done as described in Abdul-Rahman’s paper[1].The semantic di?erence is then used to adjust each future recommendation.To complete the example,remember that user A determined that user B deserves the trust category of trustworthy. If user C has determined(from previous experiences)that when user A rec-ommends trustworthy,C’s personal experience has shown that a untrustworthy experience usually occurs.In this case the semantic di?erence says to reduce A’s recommendation by one category.Therefore,C would adjust A’s trustworthy recommendation to that of untrustworthy.

6.2Trust Implementation

We propose a novel application of Role-based Trust-management language(RT0) [22]to implement the social trust model described above.RT0uses credentials to delegate trust roles from one entity to another.Determining if an entity can have a particular role relies on?nding a credential chain between the entity and the ultimate authority on that role.What follows is some background on RT0 and credential chains.

Background on RT0and Credential Chains In RT0entities(users)declare roles of the form U.r,where U is a user and r is a https://www.360docs.net/doc/771694625.html,ers can issue four types of credentials:

–Type1:U1.r←?U2

Entity U2is a member of U1’s role U1.r.U1and U2may be the same user.–Type2:U1.r1←?U2.r2

All members of U2.r2are to be included as members of U1.r1.U1and U2 may be the same users.r1and r2may be the same roles.

–Type3:U1.r1←?U1.r2.r3

Any member of U1.r2(say U2)is allowed to determine members of U1.r1by adding a credential U2.r3←?U3.

–Type4:U1.r←?f1∩f2∩...∩f k

The intersection of any number of roles and users.

As an example we present a naive,but valid,strategy for the creation of credential chains for the purpose of recommending trust.

Each user i creates a role U i.trusted.For each other user U j that U i trusts, U i issues the credential:

U i.trusted←?U j(1) In this simple case,determining if U a trusts U b is done be?nding the cre-dential chain(the number over the arrow refers back to the credential number as labeled in the paper):

←?U b

Chain:U a.trusted1

User U c can be indirectly trusted by U a by the appropriate users issuing the credentials as follows:

U a.trusted←?U b(2)

U a.trusted←?U b.trusted(3)

U b←?U c(4) The credential chain that allows U c to have the role U a.trusted is:

←?U c

Chain:U a.trusted3

←?U b.trusted4

Although this set of credentials is useful it has a draw back.All users who are directly or indirectly trusted by U b are trusted by U a.Since U a might trust U c’s data sets,but not trust U c’s recommendations,we need a more powerful set of credentials.

This example has shown the basic features of https://www.360docs.net/doc/771694625.html,ers in our system will be able to use any strategy they wish for creating roles and credential.The next section describes a better suggested strategy for creating roles and credentials.

Credential Chain Strategy We have created a strategy for creating roles and credential rules that allow for the implementation of the social trust model described in section6.1.The trust model,as presented,provides four cate-gories of trust:very trustworthy,trustworthy,untrustworthy,and very untrust-worthy.Again,these categories were chosen because of their similarity to Abdul-Rahman’s examples.However,any?nite number of categories can be chosen.

Any user i subscribing to our strategy will?rst create four basic trust roles: U i.vt:very trustworthy

U i.t:trustworthy

U i.ut:untrustworthy

U i.vut:very untrustworthy

A user j is awarded a certain amount of trust depending on which of four roles applies to that user.Credentials are needed to assign these roles to users. This set of credentials has to do with the?rst-hand experiences a user has had. These credentials require the creation of four additional roles.

U i.exp

t:Users awarded trustworthy by?rst-hand experiences U i.exp

vut:Users awarded very untrustworthy by?rst-hand experiences Because personal experience is always more important than recommenda-tions the?rst-hand experience roles will directly linked to the basic roles by the credentials:

U i.vt←?U i.exp

t(6)

U i.ut←?U i.exp

vut(8) If most of U i’s?rst-hand experiences with U j are good experiences,U i will create a credential rule U i.exp

t←?U j

Next,credentials need to be created to incorporate second hand recommen-dation of other users.If the other user subscribes to this strategy,s/he will record his or her?rst-hand experiences and create credentials according to these experiences..A user will link to his or her?rst-hand experience roles in a manner consistent with the trust model.In the model,a user must record recommen-dations of other users and compare these recommendations with his or her own ?rst-hand experiences.The di?erence between the recommended values and the observed values will be applied to all new recommendations as an adjustment. The e?ect on credential will be that a recommendation by U j of role U j.t may be,in U i’s eyes,equivalent to U i.ut.This will be the case when U j rates others higher than U i,possible because his or her standards are lower.U i may adjust U j’s recommendations by submitting the credentials:

U i.t←?U j.exp

t(10)

U i.vut←?U j.exp

vut(12) If U j had?rst-hand experiences with U k which produced the credential U j.exp

t←?U k

In this case U i has determined that U j usually recommends at one level higher than U i’s personal experience shows to be true.All of the recommendations have been adjusted down by one level.Notice that U i will not except any of U j’s recommendation to the role U i.vt.In general,the transformation from U j’s recommendations to U i’s trust values does not have to adjust all levels in the same direction or by the same about.As an example.U i’s experience with U k may l produce the credentials:

U i.vt←?U k.exp

t(14)

U i.ut←?U k.exp

vt(16) This situation probably would not happen,but is still acceptable.

If there are a signi?cant number of users making recommendations,there may be con?icting results of the credential chains(more than one basic role may be applied to a single user).For this reason the?nal decision on the appropriate role to apply to the user is made by counting the number of times each role is applied.In a similar fashion to the model,the role that was applied most is chosen.A user may even weight recommendation to achieve a weighted sum.

For the trust model to work each user should follow this strategy for creating credentials based on the semantic di?erences between his or her own experiences and the recommendations of others.However,if any user accidentally or mali-ciously creates faulty credential chains,the semantic di?erences applied to that user will adjust the recommendations accordingly.

There are many other possible strategies using RT0and credential chains. We plan on developing more and studying how di?erent strategies interact with each other.

7Future Work

We have many ideas for increasing the capabilities of our system.First,we would like to look at how much of the credential creation can be automated.Currently, validation of data sets must be done manually and rating of?rst-hand experi-ences must be done by a human.We think that some decisions about experiences can be automated and the credential chains can be updated accordingly.

Second,we would like to look at di?erent strategies that users may use in determining trust and creating credential chains.We expect to?nd that not all strategies work well together and would like to answer these questions:Which strategies do work together?Is there a best strategy?If so,what is the best strategy?

Li[22]has proposed algorithms for distributed credential chain discovery. We would like to extend Li’s work by discovering not just credential chains,but also directed credential graphs.It may be that a user trusts data using several

di?erent credential chains that form a directed graph.This graph could be used to?nd the chain that provides the greatest amount of trust.

We would also like to?nd credential chains or graphs with which we can?nd data that is trusted by some set of users.This could be used by the community to?nd the data sets which the community as a whole tends to trust.This data would be the best to use when drawing results to be presented to the community. 8Conclusion

This paper proposes an e-notebook data sharing middleware for scienti?c collab-oration.The aim of the system is to create a virtual community where scientists sharing?les are accountable for the?les they share.We would also like to encour-age the formation of natural trust views among these scientists.Accountability for shared data and the repercussions of obtaining a negative reputation will not only help scientists identify valid data but raise the integrity of the data in the entire system.Future research will re?ne the trust model as well as the data history with the goal of creating distributed community?le sharing systems with integrity similar to the professional peer review process in which malicious or incompetent users are exposed and there contributions are removed. Acknowledgments

This work is supported in part by a grant from the e-Enterprise Center at Purdue University,a gift from Microsoft Research,and grants from the National Science Foundation IIS0209059and IIS0242840.

References

1.Abdul-Rahman,A.,Hailes,S.:Supporting trust in virtual communities.In:Pro-

ceedings of the33rd Hawaii International Conference on Systems Sciences.(2000) 2.Annis,J.,Zhao,Y.,Vockler,J.,Wilde,M.,Kent,S.,Foster,I.:Applying chimera

virtual data concepts to cluster?nding in the sloan sky survey.(2002)

3.Beasley,M.,Datta,S.,Kogelnik,H.,Kroemer,H.,Monroe,D.:Report of the

investigation committee on the possibility of scienti?c misconduct in the work of Hendrik Sch¨o n and co-authors.Technical report(2002)

4.Bose,R.:A conceptual framework for composing and managing scienti?c data

lineage.In:4th International Conference on Scienti?c and Statistical Database Management.(2002)15–19

5.Brenner,S.E.:Errors in genome annotation.Trends in Genetics15(1999)132–133

6.Buneman,P.,Khanna,S.,Tajima,K.:Data archiving.In:Workshop on Data

Derivation and Provenance.(2002)

7.Buneman,P.,Khanna,S.,Tan,W.C.:Why and where:A characterization of data

provenance.In:International Conference on Database Theory(ICDT).(2001) 8.Cavanaugh,R.,Graham,G.E.:Apples and apple-shaped oranges:Equivalence of

data returned on subsequent queries with provenance information.In:Workshop on Data Derivation and Provenance.(2002)

9.Devos,D.,Valencia,A.:Intrinsic errors in genome annotation.Trends in Genetics

17(2001)429–431

10.Foster,I.,Vockler,J.,Wilde,M.,Zhao,Y.:Chimera:A virtual data system for

representing,querying,and automating data derivation.In:Proceedings of the 14th International Conference on Scienti?c and Statistical Database Management.

(2002)

11.Foster,I.,Vockler,J.,Wilde,M.,Zhao,Y.:The virtual data grid:A new model

and architecture for data-intensive collaboration.In:Proceedings of the2003CIDR Conference.(2003)

12.Frew,J.,Bose,R.:Earth system science workbench:A data management infras-

tructure for earth science products.In:SSDBM2001Thirteenth International Conference on Scienti?c and Statistical Database Management.(2001)180–189 13.Galperin,M.,Koonin,E.:Sources of systematic error in functional annotation of

genomes:domain rearrangement,non-orthologous gene displacement and operon disruption.Silico Biol.1(1998)55–67

14.Gertz,M.:Data annotation in collaborative research environments.In:Workshop

on Data Derivation and Provenance.(2002)

15.R.Gilks,W.,Audit,B.,Angelis,D.D.,Tsoka,S.,Ouzounis,C.A.:Modeling the

percolation of annotation errors in a database of protein sequences.Bioinformatics 18(2002)1641–1649

16.Greenwood,M.,Goble,C.,Stevens,R.,Zhao,J.,Addis,M.,Marvin,D.,Moreau,

L.,Oinn,T.:Provenance of e-science experiments-experience from bioinformatics.

In:Proceedings UK e-Science All Hands Meeting2003Editors.(2003)

17.Howe,B.,Maier,D.:Modeling data product generation.In:Workshop on Data

Derivation and Provenance.(2002)

18.J?sang,A.J.:The right type of trust for distributed systems.Proceedings of the

1996New Security Paradigms Workshop(1996)

19.J?sang,A.J.,Hird,S.,Faccer,E.:Simulating the e?ect of reputation systems on e-

markets.Proceedings of the First International Conference on Trust Management (2003)

20.Kaestle,G.,Shek,E.C.,Dao,S.K.:Sharing experiences from scienti?c experiments.

In:Proceedings of the11th International Conference on Scienti?c and Statistical Database Management(SSDBM).(1999)

21.Karp,P.D.:What we do not know about sequence analysis and sequence databases.

Bioinformatics14(1998)753–754

22.Li,N.,Winsborough,W.H.,Mitchell,J.C.:Distributed credential chain discovery

in trust management:extended abstract.In:ACM Conference on Computer and Communications Security.(2001)156–165

23.Mann,B.:Some data derivation and provenance issues in astronomy.In:Workshop

on Data Derivation and Provenance.(2002)

24.Patton,M.A.,J?sang,A.:Technologies for trust in electronic commerce.Electronic

Commerce Research Journal4(2004)

25.Stevens,R.D.,Robinson,A.J.,Goble,C.A.:my Grid:Personalized bioinformatics

on the information grid.Bioinformatics19(2003)i302–i304

26.Zhong,Y.,Lu,Y.,Bhargava,B.:Dynamic trust production based on interaction

sequence.Technical Report CSD TR03-006,Department of Computer Sciences, Purdue University(2003)

27.Zhao,J.,Goble,C.,Greenwood,M.,Wroe,C.,Stevens,R.:Annotating,linking

and browsing provenance logs for e-science.In:Proceedings of the Workshop on Semantic Web Technologies for Searching and Retrieving Scienti?c Data.(2003)

平台数据库及中间件招标技术要求参考

平台数据库及中间件招标技术要求参考 1.总体要求 本次采购的数据库系统和中间件软件应具备如下基本特性: 1.1安全性:保证系统数据处理的一致性,保证数据不被非法盗用和修改伪造,保证数据不因意外情况丢失和损坏,提供多种安全检查审计手段。 1.2准确性:保证系统数据处理的准确性,提供多种核查、审计手段。 1.3可靠性:保证系统可靠连续运行。 1.4可伸缩性:系统应能适应不同规模的业务,系统硬件平台和数据库应具有良好的可扩充扩展性能。 1.5可扩展性:采用组件化设计原则,以使系统能够适应将来可能出现的一些变化,新增功能时不应需要改造原软件系统。 1.6开放性:系统应采用主流的、开放的技术,以保证系统对各种数据业务的服务,以及与相关系统的互连能力。 1.7可移植性:系统还应具有较强的可移植性、可重用性,保证在将来发展中迅速采用最新出现的技术、长期保持系统的先进。 1.8实时性:实时完成大容量数据处理,对业务提供并发处理支持。 1.9易用性:应具有良好的中文操作界面、详细的帮助信息,系统参数的维护与管理通过操作界面完成。 1.10可管理性:应具有良好的管理手段,可管理安全、网络、服务器、操作系统、数据库及应用等。 1.11系统必须能够7X24小时运行,支持基于集群的部署结构。 1.12系统应具备良好的备份/恢复机制。 1.13其他:为确保产品的合法来源及售后服务的技术保障,本次投标必须提供数据库和中间件软件的原厂授权书。 1.14供应商需提供现场安装、调试,并在原厂售后服务基础上提供一年的免费现场技术指导。 2.应用服务器中间件软件的详细技术要求如下 2.1支持多协议与服务管理。 2.2支持同步数据集成服务。

应用中间件要求

投标方必须保证本项目所需软件产品获得生产厂家的合法授权,且为最新版本,并在售后服务承诺中保证提供至少一年的免费升级服务和技术支持服务。主要的应用支撑软件要求如下: 一、数据库系统 投标方提供的数据库管理系统需满足以下具体技术要求: 1、基本功能 提供丰富的数据类型支持,提供丰富的内置函数,主要包括:数学函数、字符串函数、日期时间函数、聚集函数、大对象函数等。支持自定义存储过程/函数,支持触发器,支持视图。支持完整性约束,支持事务的4种隔离级别。 支持海量数据存储和管理,数据存储量为32T以上,单个大对象的最大容量要支持到4GB。并发控制支持表锁、行锁和页锁,具有大规模并发处理能力。 支持集中的数据库管理,提供远程跨平台数据库管理工具;提供良好的性能监控、调整手段;提供跨库、跨系统数据管理能力。 2、安全要求 支持强用户身份鉴别:为用户身份鉴别提供口令、指纹和Radius等多种身份鉴别方式,并允许系统管理员自行配置用户身份鉴别类型。 支持自主访问控制机制:利用对象的ACL列表来检查某个用户是否具有对某个对象的某种访问权限,支持强制访问控制机制:提供基于标签的访问控制方式。提供多种加密方式来保证数据存储安全,至少支持外部密钥加密套件和透明加密两种方式。提供基于证书机制的数据加密传输。提供独立的安全审计,支持系统特权审计、用户审计、语句审计和对象审计四种类型的审计,既可以审计执行成功的语句也可以审计执行失败的语句。支持三权分立的安全体系,建立系统管理员、系统审计员、系统安全员的三权分立安全模型,并将访问控制的粒度细化到行级。 3、性能要求 支持多种索引,支持多种查询优化策略,支持存储过程优化、基于代价的查询优化、基于规划的查询优化,支持高效的自动数据压缩。支持物化视图,提供并行查询能力。支持一级及二级水平分区,包括:hash分区,range分区和list

数据库及中间件采购需求

数据库及中间件采购需求 中间件:ORACLE WebLogic Server Standard Edition应用服务器中间件标准版25用户数量:1套 数据库:ORACLE Database Enterprise Edition数据库企业版25用户数量:1套 其他要求: 一、合同价格 1.1本合同价格包括货物金额以及依约在交付后所需承担的售后服务价格的总和,且为完税后价格。乙方免费赠送两套最新版原厂光盘介质 二、支付和结算方式 2.1、双方因本合同发生的一切费用均以人民币结算及支付。 2.2、双方的帐户名称、开户银行及帐号以本合同提供的为准。 合同付款方式变更如下: (1)预付款:自合同签订起10个工作日内甲方向乙方预付合同总额的30%。 (2)到货款:全部货物到达合同指定现场并安装运行,甲方签字验收后10个工作日内甲方向乙方支付合同总额的30%。 (3)终验款:在完成相关集成工作并试运行/开发初验或者到货3个月后,双方签署终验证书后10个工作日内甲方向乙方支付合同总额的35%。 (4)尾款:免费保修期满一年后,对保修和维护工作进行验收,双方签署验收证书后10个工作日内甲方向乙方支付合同总额的5%。(保修期 为一年,自全部货物验收合格,双方签定最终验收报告之日起计算). 2.3、如乙方根据本合同规定有责任向甲方支付违约金或其它赔偿时,甲方有权直接从上述付款中扣除该等款项并于事后通知乙方,该情形下应当视为甲方已经依约履行了合同义务,而所扣乙方的款项金额未达到乙方依照其责任所应当向甲方支付的金额时,乙方仍应向甲方补足。同时,若乙方对甲方的扣款有异议而不能协商解决时有权依照本合同关于解决争议的约定方式解决。但,存在或解决

中间件在物联网中的应用

中间件在物联网中的应用 ——董永朋杨勋物联网是战略性新兴产业的重要组成部分,其在农业领域已有一些探索和 初步应用。农业物联网的发展,展方式的革命性变革。但与物联网技术在工业控制和电子商务等领域发展相比,农业物联网还正处于起步阶段,迫切需要对 其进行深入探索和研究,促进农业物联网技术、产品的研发和生产,改造传统 农业的现代化水平。鉴此,编辑部将近期与农业物联网相关的文章集中到一起,以期为农业物联网的相关研究人员提供参考。 物联网( The Internet of Things) 是通过无线射频识别( RFID) 装置、红外感应器、全球定位系统、激光扫描器等信息传感设备,按约定的协议,把任 何物品与互联网相连接,进行信息交换和通信,以实现智能化识别、定位、跟踪、监控和管理的一种网络。物联网改变了人与自然界的交互方式,实现人与人、人与物、物与物之间的互联,把虚拟的信息世界与现实的物理世界链接起来,融为一体,扩展了现有网络的功能和人类认识改造世界的能力。其概念是 继计算机、互联网与移动通信网之后的又一次信息产业浪潮,是一个全新的技 术领域。 1、我国物联网发展现状 我国对物联网发展高度重视,《国家中长期科学与技术发展规划( 2006 -2020 年) 》和“新一代宽带移动无线通信网”重大专项中均将物联网列入重点研究领域。《国务院关于加快培育和发展战略性新兴产业的决定》( 2009 年) 已将包括物联网在内的新一代信息技术产业被列为战略性新兴产业。2009 年以来,温家宝总理多次强调要“着力突破物联网的关键技术,及早部署后 IP 时代相 关技术研发,使信息网络产业成为推动产业升级、迈向信息社会的‘发动机’”。我国将物联网作为推荐产业信息化进程的重要策略,在实际发展中落 实物联网于各个产业中的应用。农业作为关系国计民生的基础产业,其信息化 的程度尤其受到重视。我国发展现代农业面临着资源紧缺与生态环境恶化的双

weblogic中间件介绍

w e b l o g i c中间件介绍 Company Document number:WTUT-WT88Y-W8BBGB-BWYTT-19998

目录

一、Weblogic11g概述

编写目的 ■金税三期以后的综税的产品线中间件由原来Weblogic814,全面升级为Weblogic11g,JDK统一使用及以上版本。 ■为了满足三期后运维要,全面提高运维工程师运维能力。本文档全面介绍了Weblogic11g中间件的基础操作。 功能简介 ■支持最新的 Java 平台、企业版 (Java EE) 规范及Web 服务标准,从而可简化开发并 增强互操作性,以支持面向服务的体系结构 (SOA)。 ■领先的可靠性、可用性、可扩展性和业界领先的性能。 主要优势 ■J2EE应用服务器性能记录的保持者 ■应用程序和服务的可用性和运行时间 ■更好地监视和管理生产应用程序 ■更快、更高效的开发-部署-调试周期 ■卓越的最终用户客户端可用性 ■高效快速的服务器管理 ■简化新应用程序和服务的开发 适用范围

■J2EE应用服务器 ■BS三层架构的应用服务器 Weblogic11G新特性 自调优的企业级内核 ?静态的线程池参数可以不进行设置 ?系统自动维护线程池的大小 ?自动记录系统历史的吞吐量和性能统计 ?为了达到资源的最优分配,自动优化服务器 ?没有本地代码 过载保护 ?合理的处理过量的服务–过载保护 ?根据内存与队列容量的极限值的设定拒绝请求 ?通过降低非关键业务系统的使用资源,来保证关键业务系统的正常 ?过载的时候拒绝新的请求而不是降低整个服务器的服务质量 ?优雅的意外处理 ?可以选择当发生死锁、内存溢出等关键错误时,关闭或暂停服务器动态的配置变化 ?事务式的配置变化– all or nothing! ?大部分的变化不需要重启服务器

第一章:中间件技术介绍

第一章:中间件技术介绍 1.1两层结构与三层结构 长期以来,我们一直使用着"客户端/服务器"的两层结构,这种两层的结构曾让无数人 为之兴奋和惊叹,即客户端提供用户界面、处理业务逻辑,数据库服务器接受客户端SQL 语句并对数据库进行查询,更新等操作,然后操作结果返回给客户端,如图所示。 在一个比较简单的计算机应用系统中,采用两层体系结构的确给人们带来了相当的灵活性。但随着计算机应用水平的飞速发展、企业信息化水平的不断深入、企业客户的不断增 加,以及新业务的不断出现,越来越多的用户对计算机应用系统提出了更高的要求: 1.要能够同时支持成千上万乃至更多用户的并发服务请求 2.由单一的局域网向跨多个网络协议的广域网扩展 3.不仅要支持一般的信息管理,而且还要支持关键业务的联机交易处理 4.从支持单一的系统平台和数据源转向支持异构的多系统平台和多数据源 面对用户的新需求,二层结构的应用模式由于采用客户机与服务器直接联接的方式形成了其固有的一些缺陷: 1.难以维护 clie nt/server 结构用户界面、业务逻辑和数据逻辑相互交错,通常在第一次部署的时候比较 容易,但难于升级或改进,而且经常基于某种专有的协议(通常是某种数据库协议)。它使得重 用业务逻辑和界面逻辑变得非常困难。 2 ?难以扩展 随着系统的升级,系统复杂程度大大增加,难以扩展,另外它是一个封闭的系统,很难与其他的应用系统实现互操作。 3.安全性差 客户端程序可以直接访问数据库,可通过编程语言或数据库提供的工具直接对数据库进行操作,不安全

4?性能不好 客户端直接与数据库建立连接,当有大量的并发用户存在时,会使数据库不堪重负,性能迅速下降,甚至当机。 三层结构 为解决传统二层模式与应用需求日益突出的矛盾,以交易中间件为基础框架的三层应用模式应运而生,三层结构以中间层管理大量的客户端并为其联接、集成多种异构的服务器平台,通过有效的组织和管理,在极为宽广的范围内将客户机与服务器进行高效组合。同时中间件开创的以负载平衡、动态伸缩等功能为代表的管理模式,已被广泛证实为建立关键业务应用系统的最佳环境,使在二层模式下不可能实现的应用成为可能,并为应用提供了充分的扩展余地。这种模式的成功应用已为许多国际大型企业在应用的开发和部署方面节省了大量的时间和金钱。由此促使越来越多的系统开发商和用户采用三层结构模式开发和实施其应用。 三层客户机/服务器模式的核心概念是利用中间件将应用的用户界面、业务逻辑和数据逻辑 分为三个不同的处理层,如图所示? 1.表示层(用户界面):它的主要功能是实现用户交互和数据表示,为以后的处理收集数据, 向第二层的业务逻辑请求调用核心服务处理,并显示处理结果。这一层通常采用VB, PB DELPHI等语言编写,或采用浏览器实现 2.中间层(业务逻辑):实现整个应用系统核心业务逻辑,通常把业务逻辑划分成一个个独立 的模块,用中间件提供的API结合数据库提供的编程接口实现。客户端通过调用这些模块 实现相应的业务操作。 3.数据层(数据逻辑):数据库负责管理整个应用系统的数据资源,完成数据操作。中间层上应用程序 在处理客户端的请求时,通常要存取数据库。 随着市场竞争的日益加剧和企业电子信息化建设的不断深入,高度灵活、能快速部署新服务和新应用的三层结构应用系统将成为企业信息化的必由之路。采用以中间件为基础的三层结构来架构的应用系统不但具备了大型机系统稳定、安全和处理能力高等特性,同时拥有开放式系统成本低、可扩展性强、开发周期短等优点。可以很好解决两层结构所面临的问题。中间件作为构造三层结构应用系统的基础平台,在三层结构中起着关键的作用,下一节我们将对中间件技术做一个概括性的介绍。 1. 2 中间件技术简介

智慧城市物联网中间件平台

智慧城市物联网中间件平台 采购需求文档 一、项目背景 物联网是通过信息传感设备,按约定的协议实现人与人、人与物、物与物之间的全面互联的网络,其主要特征是通过信息传感设备等方式获取物理世界的各种信息,结合互联网、通信网等网络进行信息传送与交互,采用智能计算技术对信息进行分析处理。从而提高对物质世界的感知能力,实现智能化的决策和控制。作为新一代信息技术的典型代表,与云计算、大数据等新兴热点技术并称为“智慧城市”的支柱,其应用越来越多、越来越重要。 在智慧城市建设中,物联网技术已经被广泛应用到市政、交通、应急、水务、环保、食品安全等多个领域,出现了以交通诱导、灾害预警、环保监测、食品溯源等为代表的一批典型应用,并逐渐在各个领域中发挥重要作用,智慧城市物联网应用正走向产业化和规模化。智慧城市物联网的技术体系主要由感知层、网络层、数据层、平台层和应用层组成。其中,感知层和网络层相对发展比较成熟,基本上能够满足物联网产业的发展需求。当前,物联网所面临的是数据层、平台层和应用层这三个层面上的资源整合和业务创新的问题。主要体现为以下几点: ●接入的物联网硬件设备种类和数量日益增多,不同类别的设备运行环境 不同,通信协议也不同,而上层应用需要对这些这些设备进行统一管理, 包括信息获取和设备控制。这需要应用的支撑平台可以适配各种异构环 境,并且有接入海量硬件设备的能力; ●城市级的应用需要接入海量的物联网设备,海量设备会产生大量的并发 事件和传感数据,物联网应用需要处理大量的并发操作和数据存储。这 需要应用的支撑平台能够提供大量的计算和存储能力,使用云计算技术 是目前的主要方式。 ●智慧城市建设涉及到市政、交通、能源、教育、医疗等各个领域,不同

什么是地图发布中间件及其功能应用介绍

什么是地图发布中间件及其功能应用介绍 一、海量影像地图数据发布首选——中间件 如果需要发布海量影像数据快速构建全国离线二维GIS地理信息系统或全球离线三维地球触摸GIS系统,则需要由硬件、软件、数据和GIS平台四部分组成。 1)硬件 硬件主要包括地图数据服务器和客户端PC机。 服务器:主要用于安装中间件、布署GIS应用平台和存储全国卫星影像数据。 客户端:用于加载GIS平台,并接收中间件发布的影像数据、地名路网数据和高程数据。 2)软件 软件主要包括《水经注地图发布服务中间件》(简称“中间件”)和《水经注万能地图下载器》(简称“下载器”)。 中间件:用于发布全国或全球海量卫星影像数据、地名路网和高程数据。 下载器:用于下载卫星影像数据、地名路网和高程数据。 3)数据 用户可以自行下载数据或直接购买下载好的数据。 自行下载:卫星影像数据、地名路网数据和高程数据可以用《水经注万能地图下载器》自行下载。 直接购买:购买之后,会通过邮寄硬盘(全国数据)或阵列柜(全球数据)

的方式为用户提供。 4)GIS平台 由于中间件只是一个基于URL请求返回瓦片数据的功能部件,因此只要可以支持瓦片式影像加载的GIS平台都可以进行调用。 这里推荐几个GIS开发平台供选择: 1)Google Map 离线API 2)Openlayers 二维开源平台 3)ArcGIS API for JavaScript 4)Cesium 开源三维地球平台 5)OsgEarth开源三维地球平台 二、什么是地图发布中间件 简单的讲,地图发布中间件就是为客户端提供影像瓦片的一个Windows系统服务。它只做一件事,也就是客户端通过URL请求的方式,可以快速返回影像瓦片、地名路网瓦片和高程瓦片数据。 获取影像URL示例 http://127.0.0.1:8080/getImage?z=6&y=62&x=35 获取地图路网URL示例 http://127.0.0.1:8080/getlabel?z=6&y=62&x=35 获取高程URL示例 http://127.0.0.1:8080/getDem?z=6&y=62&x=35

中间件安装步骤

中间件服务器安装步骤 1、安装WINDOWS 2003 32位操作系统,更改计算机名,设置许可证每服务器可连接600 个客户端。 2、在控制面板里的增加删除程序界面,如图安装 3、安装相应驱动程序。 4、设置IP,并加入域https://www.360docs.net/doc/771694625.html,。 5、先安装R2再安装操作系统的补丁SP2。 6、安装.NET FRAMEWORK 2.0,安装语言包补丁,安装.NET FRAMEWORK 2.0 SDK。 7、安装https://www.360docs.net/doc/771694625.html,。 8、安装ORACLE客户端,要求安装ODPNET版本第一项和第五项。 9、安装.NET FRAMEWORK 32位补丁1。 10、配置COM+组件服务,如图

COM 安全中四个按钮都要打开,并加入EVERYONE 用户都有全部权限。

11、 把ICARECONTROLLER 的域用户添加到本地管理员组中 12、 拷贝更新服务器上最新的DEBUG 文件夹 13、 运行DEBUG 目录下“生成三层中间件服务器目录.bat ”文件,注册COM+组件

14、修改ICARESYSTEM的属性,如图

15、 把ICAREDA TA.DLL 和IBM.Data.DB2.dll 放到GAC 里,操作方法如下: 建立批处理文件,内容如下:建立bat 文件,内容包括如下,然后把它放在

"D:\Debug\gacutil" /il RegGacFileList.txt pause 然后新建RegGacFileList.txt,里面包括ICAREDATA.DLL、IBM.Data.DB2.dll、EMR_ICAREDATA.DLL和EMR_BASE_VO.DLL文件名即可,格式如下: iCareData.dll IBM.Data.DB2.dll EMR_ICAREDATA.DLL EMR_BASE_VO.DLL 也可以通过控制面板下管理工具中的.NET FRAMEWORK2.0工具进行添加上述四个文件。 16、更新中间件时 a)必须先右键关闭ICARESYSTEM三次,并禁用 b)然后反注册 c)再拷贝新文件 d)再注册 e)启动服务,完成之后再检查ICARESYSTEM的属性,确保没有变成库模式。 17、如果某个中间件之前已经放在GAC里面,首先必须在GAC里进行反注册该中间 件即可,操作方法如下: 建立批处理文件,内容如下 "D:\Debug\gacutil" /ul UnRegGacFileList.txt pause 然后新建UnRegGacFileList.txt,里面包括需要反注册的文件名即可,格式请参考如下:HISReg_SVC(特别注明:不需要加文件扩展名) 18、配置自动更新服务: a)必须先右键关闭UpdateSystem_Svr三次,并禁用 b)再拷贝新文件 c)修改UpdateFiles.xml的版本号和需要更新文件的标志 d)启动服务UpdateSystem_Svr 19、客户端只需要运行icare.exe即可自动更新。

数据库中间件使用场景分析

数据库中间件使用场景分析数据库场景比较 PS:涉及到金钱方面的事务处理,建议使用Oracle。 数据库优点缺点场景 Oracle 基本适合所有业务维护成本和License成 本高 电信,电力、银行、支付以及涉及到金钱 方面等综合性企业。(事务型) MySQL 结构简单,部署方便,社区 成熟,稳定性非常好, 良好的事务和SQL支持 扩展性差,软件本身性 能瓶颈大, 没有成熟的集群方案。 Schema复制。 百亿以内的数据存储, 对数据安全性和事务支持有要求。主要存 储对数据状态有要求和更新频繁的数据。 (事务型) MongoDB Schema--free,快速开发, 本身支持集群如sharding, 支持空间索引等; 锁的粒度大,并发性能 差,性能受限于内存, 解决方案有待考验。 1.LBS(基于位置服务;地理坐标,或大地坐 标),缓存,小文件存储。 2.CMS内容管理系统; 3.社交网络图数据库设计. 4.MongoDB主要用于存储计费数据、日志 数据和流水数据 Hbase 基于Hadoop生态系统,良 好的扩展性,高写入能力。 数据自动分片。 架构复杂,维护成本 高。 搜索,数据写入非常高,监控数据。 1.典型互联网搜索问题 2.捕获增量数据 3.内容服务 4.信息交换 HBase主要用来做数据分析和存储大数据 内容。 Redis 高性能,部署简单,非常的 数据类型支持, 支持数据持久化,集群方案 支持。 性能受限于内存,单进 程问题。 适合小数据高读写场景。缓存服务。 1.保存点击数据(计数器) 2.在哈希表中保存用户信息 3.用集合保存社交网站圈子数据

MySQL还是PostgreSQL? 1、如果你的应用对数据的完整性和严肃性要求不高,但是追求处理的高速度。例如是一个论坛和社区,你应该使用MySQL。 2、你的应用是一个严肃的商业应用,对数据完整性要求很高。而且你希望对一些商业数据逻辑进行很好的封装,例如是一个网上银行,你应该使用PostgreSQL。 3、你的应用处理的是地理数据,由于R-TREES的存在,你应该使用PostgreSQL。 4、等等 从Oracle转向MySQL主要是出于三个方面的原因: 第一,降低运维成本。Oracle数据库自动化运维实现难度和成本较高,而MySQL运维自动化难度和成本相对较低,当数据库实例不断成倍增长的时候,使用MySQL可以在有限人力的情况下维护更多的数据库实例。 第二,降低软件成本。Oracle License成本较高,MySQL及其分支目前是免费的。 第三,提高可扩展性。MySQL是开源数据库,便于有技术能力的公司根据业务发展情况自己开发定制一些数据库周边服务,使数据库使用的扩展性提高,而Oracle对这方面的支持比较一般。 Hbase场景说明 捕获增量数据 数据通常是细水长流,累加到已有数据库以备将来使用,例如分析,处理和服务。许多HBase使用场景属于这个类别——使用HBase作为数据存储,捕获来自于各种数据源的增量数据。例如,这种数据源可能是网页爬虫,可能是记录用户看了什么广告和多长时间的广告效果数据,也可能是记录各种参数的时间序列数据。我们讨论几个成功的使用场景和公司。 1.捕获监控参数

中间件开发方案

中间件开发方案 一使用中间件原因 由于呼叫中心提供的调用呼叫中心的方法只支持单独页面的独立调用,无法满足BS架构的用户在多页面调用呼叫中心功能,因此需要使用中间件服务器连接呼叫中心系统,模拟建立独立通信通道,将多页面调用所需的功能发送给中间件,由中间件作为呼叫中心的唯一调用源,以此保证在符合呼叫中心调用机制的情况下完成BS架构的呼叫中心完整功能。 二开发方案1(CS架构客户端调用方式) 中间件组成部分:(服务器端和客户端) 1、服务器端功能: a)CRM用户和呼叫中心坐席关联,记录在中间件坐席信息表中。 b)接收和记录客户端状态和客户端传来的坐席状态。 c)接收和记录BS架构发送的呼叫中心调用命令。 d)记录客户端和BS架构调用呼叫中心的调用记录及调用结果。 2、客户端功能: a)连接呼叫中心的服务器,实现坐席登录功能。 b)获取呼叫中心服务器上的坐席状态,发送给中间件服务器端。 c)用呼叫中心提供的CS开发文档和开发ocx、dll调用呼叫中心的话务功能。 d)调用呼叫中心话务功能的时候讲调用记录发送给中间件服务器端。 e)接收中间件服务器端传来的BS架构调用呼叫中心话务功能的消息,并根据消 息判断触发呼叫中心话务功能类型及参数,翻译后发送给呼叫中心服务器。 f)在客户端转接、强插、监听等功能调用的时候读取中间件服务器端坐席状态表, 获取在线空闲坐席信息、在线通话中坐席信息等列表,并根据列表中的信息整 理成为调用参数,单击或双击列表中坐席调用呼叫中心话务功能。 g)监控来电事件,根据来电号码、客户端登录坐席,实现CRM弹屏。 3、CRM话务功能调用: a)发送命令消息给中间件服务器。500ms后查询服务器执行结果,若执行结果为 失败则显示失败消息,若执行结果为成功则无动作。 b)读取中间件服务器端坐席状态表,获取在线空闲坐席信息、在线通话中坐席信 息等列表,并根据列表中的信息整理成为调用参数,单击或双击列表中坐席将 转接命令消息发送给中间件服务器,并实现销售线索和来电客户数据的自动生 成或关联。

数据库中间件高级技术支持服务说明

数据库中间件高级技术支持 服务说明 -标准化文件发布号:(9456-EUATWK-MWUB-WUNN-INNUL-DDQTY-KII

数据库和中间件技术支持 服务说明书 版本号:V2.0 赛尔网络 2010年4月28日

目录 第一章、公司简介 (5) 第二章、服务内容提要 (7) 第三章、数据库和中间件技术支持与服务详述 (9) 3.1、远程支持服务 (9) 3.1.1、中文电话支持服务(7*24小时) (9) 3.1.2、Email服务(7*24小时) (9) 3.2.现场服务 (10) 3.2.1.数据库、中间件安装调试服务 (10) 3.2.1.1、数据库安装调试服务 (10) 3.2.1.2、中间件安装调试服务 (11) 3.2.2定期系统健康检查服务 (12) 3.2.2.1、数据库方面的健康巡检 (12) 3.2.2.2、中间件方面的健康巡检 (13) 3.2.2.3、操作系统方面的检查 (15) 3.2.3.性能优化服务 (16) 3.2.4.数据库备份恢复策略的制定和测试服务 (18) 3.2.5.数据库和中间件升级及迁移服务 (20) 3.2.6.数据库和中间件应急服务 (20) 3.2.7.重大事件待命服务 (21) 3.2.8、制定数据库和中间件管理规范服务 (21) 3.2.9.其他现场服务 (22) 第四章服务质量保证及验收标准 (23) 4.1、服务项目的组织结构及人员安排 (23) 4.2、客户服务档案 (24) 4.3、服务效果的验收 (25) 第五章赛尔网络服务质量保障体系 (26)

5.1 技术支持总体流程 (26) 5.2 现场支持工作流程 (27) 5.3 健康巡检工作流程............................................................................ 错误!未定义书签。 5.4 重大紧急事件处理流程 (29)

中间件介绍

中间件介绍 文档编制序号:[KKIDT-LLE0828-LLETD298-POI08]

中间件介绍 1、Ice: ICE(Internet Communications Engine)是ZeroC提供的一款高性能的中间件,基于ICE可以实现电信级的解决方案。在设计网站架构的时候可以使用ICE实现对网站应用的基础对象操作,将基础对象操作和数据库操作封装在这一层,在业务逻辑层以及表现层(java,php,,python)进行更丰富的表现与操作,从而实现比较好的架构。基于ICE的数据层可以在未来方便的进行扩展。ICE支持分布式的部署管理,消息中间件,以及网格计算等等。 Zeroc推出的一种分布式的面向对象中间件,解决分布式的异构计算。可以用 C++,Java,c#等进行分布式的交互计算。 主要设计目标是: ·成为适用于异种环境的平台。 ·具有一组完整的特性,支持广泛的领域中的实际的的开发。 ·去掉不必要的复杂性,使平台更易于学习和使用。 ·是一种在、内存使用和CPU开销方面都很高效的实现。 ·是一种具有内建安全性的实现,使它适用于不安全的公共网络。 2、JBoss: 是一个基于J2EE的的。 JBoss代码遵循LGPL许可,可以在任何商业应用中免费使用,而不用支付费用。JBoss是一个管理EJB的容器和服务器,支持EJB 、EJB 和EJB3的规范。但JBoss核心服务不包括支持servlet/JSP的WEB容器,一般与Tomcat或Jetty绑定使用。 在J2EE领域,JBoss是发展最为迅速的应用服务器。由于JBoss遵循商业友好的LGPL授权分发,并且由开源社区开发,这使得JBoss广为流行。 另外,JBoss应用服务器还具有许多优秀的特质。 JBoss运行后后台管理界面 其一,将具有革命性的JMX服务作为其; 其二,本身就是(Service-Oriented Architecture,); 其三,具有统一的类装载器,从而能够实现应用的和热卸载能力。 因此,高度模块化的和松耦合。JBoss应用服务器是健壮的、高质量的,而且还具有良好的性能。 1、JBoss是免费的,J2EE的实现,通过许可证进行发布。但同时也有的,开源和闭源流入流出的不是同一途径。 2、JBoss需要的内存和硬盘空间比较小。 3、安装便捷:解压后,只需配置一些即可。 4、JBoss支持"热部署",部署BEAN时,只拷贝BEAN的文件到部署路径下即可自动加载;如果有改动,也会自动更新。 5、JBoss与Web服务器在同一个中运行,Servlet调用EJB不经过网络,从而大大提高运行效率,提升安全性能。

数据库中间件及其几种技术比较

数据库中间件及其几种技术比较 摘要:本文阐述了数据库中间件的概念,功能,原理,介绍了现今数据库中间件的几种主要技术,并进行了比较。 关键字:数据库中间件 1、数据库中间件的基本概念 数据库中间件是处于底层数据库和用户应用系统之间的,主要用于屏蔽异构数据库的底层细节问题的中间件,是客户与后台的数据库之间进行通讯的桥梁。当客户向Web Server发出对某个数据库的SQL请求时,通过数据库中间件搜索匹配的数据库连接,并将SQL请求转发给对应的数据库服务器,通过其对数据库进行操作。 数据库中间件的主要功能:(1)支持常用大型数据库的各种操作。如ORACLE ,DB2, MYSQL等常用数据库。(2)提供统一接口, 屏蔽数据库之间的操作差异。(3)封装复杂烦琐的数据库应用接口和数据库操作过程,简化应用程序的数据库操作, 提高应用程序开发效率。(4)支持常用的操作系统。如Windows、UNIX、Linux 等,便于应用代码在各平台之间的移植。(5)支持多线程, 可以提供多线程与线程库, 满足各种场合应用。 数据库中间件(UniWeb Server)工作原理:让其作为前端的客户与后端的数据库之间进行通信的桥梁,当客户向数据库中间件发出对某个数据库的SQL请求时数据库中间件搜索当前可用的与该数据库的连接(UniTcl Server) 通过UniTcl Server将SQL请求转发给对应的数据库服务器,数据库服务器执行SQL语句后将结果通过UniTcl Server 返回给数据库中间件,再由它返回给客户整个数据库中间件的体系结构采用的是三层(Three-tier)客户机/服务器模型,中间件与各个客户的数据通信采用流套接字(Stream Socket)机制实现并

TUXEDO中间件介绍及应用

TUXEDO中间件介绍及应用 一、前言 首先介绍一下什么是中间件?中间件是一种独立的系统软件或服务程序,分布式应用软件借助这种软件在不同的技术之间共享资源,中间件位于客户机服务器的操作系统之上,管理计算资源和网络通信。 中间件屏蔽了底层操作系统的复杂性,使程序开发人员面对一个简单而统一的开发环境,减少程序设计的复杂性,将注意力集中在自己的业务上,不必再为程序在不同系统软件上的移植而重复工作,从而大大减少了技术上的负担。 世界著名的咨询机构Standish Group在一份研究报告中归纳了中间件的十大优越性: ●缩短应用的开发周期 ●节约应用的开发成本 ●减少系统初期的建设成本●降低应用开发的失败率●保护已有的投资●简化应用集成 ●减少维护费用 ●提高应用的开发质量●保证技术进步的连续性●增强应用的生命力 Tuxedo是第一个严格意义上的中间件产品。Tuxedo是1984年在当时属于A T&T的贝尔实验室开发完成的,但Tuxedo在很长一段时期里只是实验室产品。直到BEA公司1995年收购Tuxedo后,使Tuxedo现已经发展成为交易中间件领域事实上的标准。 TUXEDO是在企业、Internet 这样的分布式运算环境中,开发和管理三层结构的客户/服务器型关键任务应用系统的强有力工具。它具备分布式事务处理和应用通信功能,并提供完善的各种服务来建立、运行和管理关键任务应用系统。开发人员能够用它建立跨多个硬件平台、数据库和操作系统的可互操作的应用系统。 二、TUXEDO的组件软件模型 TUXEDO采用三层结构的组件软件模型。 图1 BEA TUXEDO 的组件软件模型概要

WEB数据库与中间件技术解决方案

Web数据库与中间件技术 随着Internet/Intrranet的兴起与发展,Web服务器与数据库服务器的连接显得越来越重要,许多厂家不断推出新技术、新产品,使得连接更加简洁、迅速和方便。Web与数据库连接技术已成为基于Web的信息管理系统的核心,为Internet上的电子商贸打下了基础。 一般来说,通过Web页实现对数据库访问,在整个系统中关键的技术是中间件的解决方案。 中间件负责管理Web服务器和数据库服务器之间的通信并提供应用程序服务。由于驻留在Web服务器上,因而中间件软件能够调用作为Web服务器和数据库服务器间"传输机制"的外部程序或"编码",并将执行查询等以HTML页面或纯文本的形式将信息返回给最终用户。数据库服务器负责管理驻留在数据库服务器中的数据。 一、当前几种流行的中间件的解决方案 1.通用网关技术(CGI) CGI是一种Web站点上可以用来访问Web站点的用户交互的各种程序的标准,使用CGI脚本允许用户在浏览器中等服务器上的数据库交互,完成对数据库的各种操作。 几乎使用的服务器软件都支持CGI,开发者可以使用任何一种Web服务器内置语言编写CGI,包括Perl语言,C,C++,VB和Delphi等。 CGI的工作原理是浏览器通过Web页面的表单搜索参数,这些参数通过HTTP传递Web服务器,在服务器通过CGI脚本分析参数(命令行参数或环境变量),同时启动通路程序,把分析后的参数转化为SQL命令,交数据库服务器执行,然后CGI程序返回处理结果给Web服务器,最后向客户机返回HTML或纯文本格式的结果并断开连接。 CGI缺点是执行速度较慢,Web服务器每启动一个数据查询服务,就必须启动一个新的CGI 进程,相对服务器资源代价比较高。 2.ASP(Active Server Pages) ASP是一种开放的,可以将HTML脚本及可重用的Active Server组件结合在一起以建立高效的动态的基于Web的应用程序环境,利用ASP,可以增加运行在服务器端的脚本的新特性,

中间件技术知识点及习题

第10章中间件技术 10.1 基本内容分析 10.1.1 本章重要概念 (1)中间件的定义和作用 (2)ODBC分层的体系结构:应用程序,驱动程序管理器,DB驱动程序,ODBC 数据源。 (3)ODBC接口:应用程序基本流程,ODBC句柄,ODBC连接,SQL语句的执行。ODBC两套符合性级别。典型的DB应用系统开发工具。 (4)JDBC的基本功能,JDBC的结构,JDBC接口。 10.1.2 本章的重点篇幅 (1)ODBC分层的体系结构(教材P403的图10.2)。 (2)ODBC应用程序的基本流程(教材P408的图10.6)。 (3)JDBC驱动程序(教材P427-428),JDBC API接口(教材P430)。 10.2 教材中习题10的答案 10.1什么是中间件?有什么作用? 答:中间件是分布式环境中保证OS、通信协议、数据库等之间进行对话、互操作的软件系统。 中间件的作用是保证了客户和服务器间的联系,使网络、数据库、操作系统对于应用软件的开发界面透明化。 10.2 试解释中间件的三个透明性。 答:中间件的网络透明性是指中间件能支持所有类型的网络。 中间件的服务器透明性是指不管服务器上的DBMS是何种型号(ORACLE、

SYBASE、DB2等),一个好的中间件都能通过标准的SQL语言与不同DBMS上的SQL语言连接起来。 中间件的语言透明性是指客户机可用任何开发语言进行发送请求和接受回答,被调用的功能应该像语言那样也是独立的。中间件还应该保证开发语言的数据类型和服务器上数据库使用的数据类型之间能够相互转换。 10.3 ODBC技术与传统的数据库编程方式有什么区别? 答:传统的DB编程方式是“主语言+DML”,但一个应用程序却不能访问不同DB服务器上的数据。 ODBC技术实际上是一个公共接口API,使用ODBC技术,同一个应用程序就可以访问不同DB服务器上的数据。 10.4 ODBC技术有什么作用?其卓越贡献是什么? 答:ODBC技术的作用是使应用程序与DBMS在逻辑上可以分离,使应用程序具有数据库无关性。 ODBC的卓越贡献是使应用程序具有良好的互用性和可移植性,并且具备同时访问多种DBS的能力,从而克服了传统数据库应用程序的缺陷。对用户而言,ODBC驱动程序屏蔽掉了不同DBS的差异。 10.5 ODBC的体系结构有哪几层?试叙述各层之间联系。 答:ODBC的体系结构有四层:应用程序,驱动程序管理器,DBMS驱动程序、数据源。 应用程序要完成ODBC外部接口的所有工作,用C语言和ODBC函数来编应用程序。 驱动程序管理器管理应用程序和DBMS驱动程序之间的交互作用,为应用

SMO物联网中间件平台系统

TANSUOYANJIU / 探索研究 62 质量管理 SMO物联网 中间件平台系统 物联网(The Internet of Things)是指物体的信息通过各类智能感知仪器装置结合RFID技术,经过传感网和通讯网,到达指定的信息处理中心,最终实现物与物,人与物之间的自动化信息交互与处理的智能网络。物联网是继个人计算机、互联网之后的全球信息化的第三次浪潮。物联网产业链四层体系:感知设备层、网络传输层、物联网中间件和应用管理层。 本文主要讲述一款针对物联网应用终端信息进行有效采集和整合的中间件平台系统,具备强大的数据追溯和数据监控功能。系统通过SMO服务对象管理技术,将所有终端设备、业务逻辑处理等均看做是服务对象进行管理,实现快速建立应用模型;并采用分层设计的中间件软件平台软件,实现分布式实时交互业务处理逻辑的变更而不需要重新部署传感网;系统基于SOA架构多种形式的业务流程和业务逻辑处理服务调用,实现整个系统部署灵活变更和快速重构。 范小兴/文 (作者单位:福州欣创摩尔电子科技有限公司) 物联网中间件国内外研究情况 “物联网”的概念于1999年由麻省理工学院的Auto-ID实验室提出,但真正作为全球战略性产业,引起人们重视还是在2008年年底IBM公司所提出的“基于物联网技术的智慧地球”概念。其后2009年,美国总统奥巴马在就职后,为了摆脱经济危机阴影的笼罩,提出了两项新技术新能源和“物联网”,并将之作为美国在21世纪保持和夺回竞争优势的方式。 2010 年,物联网发展被正式列入国家发展战略,十二五期间,物联网将重点投资智能电网、智能交通、智能物流、智能家居、环境与安全检测、工业与自动化控制、医疗健康、精细农牧业、金融与服务业、国防军事等十大领域。各地各级政府纷纷出台物联网产业发展规划,并大力兴建物联网示范工程。作为物联网感知层关键技术的RFID 及相关产业也受到了前所未有的重视,获得了难得的发展时机。 技术要点和关键技术 SMO物联网中件间平台系统是一款针对物联网数据信息进行有效采集和整合的操作系统级中间件系统,其具备强大的数据追溯和数据监控功能,可广泛应用于制造业、食品安全追溯和公共安全等物联网应用领域。 技术要点 采用S M O 服务对象管理器平台(Service Manager of Object),将所有终端设备、业务逻辑处理等均看做是服务对象进行管理,用于监视采集终端设备运行状态、业务接口服务和设备接口服务调度管理、设备配置管理及业务服务配置等,完成物联网应用系统的传感网部署和分布式实时交互业务处理。 系统将物联网应用中最容易变化的采集技术、传输技术、控制技术和控制逻辑抽取出来,形成一个数据采集控制 系统中间件平台,使得上层应用无需关

中间件技术介绍

中间件技术介绍 中间件(middleware)是基础软件的一大类,属于可复用软件的范畴。顾名思义,中间件处于操作系统软件与用户的应用软件的中间。中间件在操作系统、网络和数据库之上,应用软件的下层,总的作用是为处于自己上层的应用软件提供运行与开发的环境,帮助用户灵活、高效地开发和集成复杂的应用软件。 在众多关于中间件的定义中,比较普遍被接受的是IDC 表述的:中间件是一种独立的系统软件或服务程序,分布式应用软件借助这种软件在不同的技术之间共享资源,中间件位于客户机服务器的操作系统之上,管理计算资源和网络通信。 IDC对中间件的定义表明,中间件是一类软件,而非一种软件;中间件不仅仅实现互连,还要实现应用之间的互操作;中间件是基于分布式处理的软件,最突出的特点是其网络通信功能。 中科院软件所研究员仲萃豪形象地把中间件定义为:平台+通信。这个定义限定了只有用于分布式系统中的此类软件才能被称为中间件,同时此定义还可以把中间件与支撑软件和实用软件区分开来。 目前,中间件发展很快,已经与操作系统、数据库并列为三大基础软件。中间件主要分为以下几类:

1.通信处理(消息)中间件 此类中间件能在不同平台之间通信,实现分布式系统中可靠的、高效的、实时的跨平台数据传输(如Tong LINK、BEAe Link、IBM的MQ Series等)。这是中间件中唯一不可缺少的,是销售额最大的中间件产品。 2.交易中间件 在分布式事务处理系统中要处理大量事务,常常在系统中要同时做上万笔事务。例如在北京市就要设置各种运载汽车,完成日常的运载,同时要随时监视汽车运行,出现故障时,要有排除措施,发生堵塞时要进行调度。在联机事务处理系统(OLTP)中,每笔事务常常要多台服务器上的程序顺序地协调完成,一旦中间发生某种故障时,不但要完成恢复工作,而且要自动切换系统,达到系统永不停机,实现高可靠性运行;同时要使大量事务在多台应用服务器能实时并发运行,并进行负载平衡地调度,实现昂贵的可靠性机和大型计算机系统同等的功能,为了实现这个目标,要求系统具有监视和调度整个系统的功能。BEA的Tuxedo由此而著名,它成为增长率最高的厂商。一个事务处理平台,根据X/OPEN的参数模型规定,应由事务处理中间件、通信处理中间件以及数据存取管理中间件三部分组成。东方通科技公司的Tong LINK和TongEASY实现了这个参考模型规定。3.数据存取管理中间件

相关文档
最新文档