Paper5

Paper Title:

Three Critical Questions

Group 1:

Member Name:Vallurupalli Anusha

1)Can the service be automated as soon as it is delivered to the customer?
2)Isn't the adoption of SOA in SCA implementation time consuming?
3)To what extent can the SCA be implemented feasibly in practice?

Member Name:Fazlun

1) The author says that the current architecture does not provide a fully fledged mechanism of providing personalized service along with contracts to all the requestors, he says that standardizing industry specific semantics, and incorporating basic infrastructure with semantic capabilities can make it possible, but the system has a lot more complexities to this, how feasible is this, can we finally reach a stage where the contracts are as rich and personalized as the requestor needs, he does not give us feasible solutions, he leaves them unaddressed.
2) The basic block of SOA is the SOAP which is built above HTTP and this provides messaging, but this at the same time has a number of prohibitive constraints to the Web developer, SOAP at the same due to its verbose XML format is slow when compared other middleware technologies, so though the author says that QOS attribute of reliable messaging can be provided using SOAP, we could get better service by sophisticating the protocol more or by putting in other protocols.
3) The author says that SCA is the most complete specification, but the SCA does not satisfy this for a large set of services, it also does not provide the feature of automatic service discovery for alternative services, so we could implement these features to make the architecture more robust.
4) Since in general the SOA environment is loosely coupled how does it support tightly distributed acid transactions if it comes across any, does it have a specification for it.
5) The ability to attach policies to specific messages is advantageous in many aspects, but what if there is an overlap, or if there is a necessity of applying two policies, how are precedence issued solved.

Group 2:

Member Name:Anuradha Pathuri

1. The architecture of SCA proposes to have implementations in different languages.
Is a single service implemented in different languages?
Or is it that each component of the service is implemented in a different language?
How would the service be resolved and understood if each component is implemented in a different language?
2. The paper proposes SCA, an architecture which proposes to implement reusable components of a service,
but if we consider one particular component of a service, it could be used in several different ways in different services.
In such a case, how can this make sense to reuse the single implementation of the component in all the services?
3. While aggregating the QoS properties of individual services,
the quality standards of two or more services may contradict with one another.
What kinds of policies are followed in such a scenario?
4. The major issue of the SOA is the visibility issue; incase if a service fails to work,
then the problem is not visible to the user of the service. The user can only see the top level services but not its subservices.
How is this issue resolved in the proposed SCA architecture?

Group 3:

Member Name: Bindu

1. As the paper presents, SOA contracts are defined in machine- readable formats since the main focus is on automating the integration and this would lead to maintaining human readable documentation separately. Is it not time consuming for the main specifications to be separately encoded in XML?
2. As stated in paper the ACID transactions use the automatic rollback recovery and the business activity assumes that the compensation operation which is defined at business level is going to carry out the error recovery. How is this error recovery done?
3. It is given in the paper that the service providers can avoid the unnecessary processing expense by having the ability to attach policies to specific messages which require them the most. How can we differentiate the ones which require encryption and digital signatures from the ones which don’t?
4. It is listed in the paper that QoS properties of components and wires are implicitly composed in ways that are difficult to understand. How can we implement QoS policies effectively if the QoS properties are difficult to understand?

Member Name: Yogendra Siva Ram, Boppana
1) Automatic integration between service components is not yet attained and a core goal to SOA vision. Even it is achieved, will the automatic integration perform as the same level as the present component integration?

2) The policies on wires between components by associating QoS properties with the binding attached to the wire are specified by the component assemblers. Are there any specific standards that define how these policies are framed? If yes, they are not mentioned clearly in the paper.

Group 4:

Member Name:Lattupalli Nikhil Reddy

Critical Questions: 1. Is XML the apt language for the interoperability? Though it is one of the leading markup languages it is still a text based semi dynamic language, can it be used as language for the communication of components in a SCA.
2. Contracts form the middle layer for providers, register and consumers. Are the contracts made flexible enough to enable any sort of change in requirements for a service i.e., if a consumer wants to modify a small change in the service, which requires the contract to extend its limitations, will it be possible for the contract to extend its limitations for certain users?
3. Is the automatic integration between service components achievable even if we use service discovery and automatic composition? yes it is very exciting to hear that it automatically integrates the service and consumer but the knowledge in integrating the components without such discovery and composition is not quiet possible. Is it?

Group 5:

Member Name:SUNIL KUMAR GARREPALLY

1. It was mentioned that SOA provides automatic integration. What methods are used to provide automatic integration? Is it possible to provide complete automatic integration without any human intervention?

2. The paper discussed about various specifications and protocols of QOS. Whether all the specifications or protocols can be used with other set of protocols for the same system. Won't they contradict each other?

3. It was mentioned that composite components are used in SOA. These components inside the composite components are connected through complex connections. So, is it easy to make changes to the components inside them and execute if there are any small changes to be made?

Group 6:

Member Name:BHARGAV SANDEEP RAMAYANAM

1. In the paper the author said that there is a service stack for storing the services. It is built on SOAP. Does the exchanging process get slower because of the XML format messages?

2. In the paper the version of Web Services Description Language the author used is 1.1. But the latest version is WSDL 2.0. Does all the specifications and policies that are mentioned in the paper are compatible with the latest version?

3. According to the author the stack supports ACID properties. By these ACID properties one of the QoS properties "ability to recovery from errors" will be solved. But these transactions are not completely compatible for SOA. Using stack that supports ACID properties is useful or not?

Group 7:

Member Name:Sandeep Kumar Yarala Padmarao

Critical Questions:

1) What is the strength of security implementation done in "QoS" specification like the use of digital signatures,encryption mechanism ? Is the entire mechanism secure up to the present standards?Security mechanisms have to be off high standard when implementing at such a high level.

2)Why structural composition model be preferred over process oriented model?To what type of service or situation a model should be preferred and what are the pros and cons of each model?

3) How is the trust issue solved? The paper does not discuss about solving this issue which is a major problem in "SOA".The paper just specifies the security standard which are implemented .For example digital signatures for message integrity,encryption for confidentiality.

Group 8:

Member Name: Nikhilesh Katakam

Critical Questions :
1)This paper claims a number of concerns for the QOS, such as the security, reliability and the ability to recover form errors, but could not specify

what methods to be followed to achieve them?
2)This paper describes two models for the service composition in SOA, Process oriented and structural oriented, which one is more feasible and why is it

feasible was not mentioned in the paper?

3)Is it not difficult to standardize the contracts , as they will be many organization who would want to use the services and each of them would have

their having their own requirements?

Group 9:

Member Name:Muppalla Durga Maheswari

1.How can we predict the properties of components and wires in service component architecture which are implicitly composed?

2. The paper states that the distributed ACID transactions are not well suited to the service oriented architecture how can we say that they are not well suited if they are supported by the web services stack.

3.How can policies directly affect service composition in service oriented architecture?

4. How the web services specification stack that provide standards for service descriptions support quality of service protocols.

Group 10:

Member Name:Jonnalagadda, Harideep reddy

1) What is SOAP's logical messaging mean????? Is the suggested use of XML encoding not difficult to be used by other systems???

2) As there are many policies which are to be understood by third parties. Can a unique policy be designed so that all policies are covered??? So that it would decrease the complexity of understanding.

3) How can the distributed acid transactions be work well in the SOA??? Is there any solution that could be designed that handles the recovery well???

Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License