1 . (10%) State the problem the paper is trying to solve.
This daily news is trying to demonstrate how Airavat, a MapReduce-based system pertaining to distributed calculations provides end-to-end confidentiality, integrity, and personal privacy guarantees utilizing a combination of necessary access control and gear privacy which gives security and privacy ensures against info leakage.
2 . (20%) Express the main contribution of the paper: solving a fresh problem, suggesting a new algorithm, or offering a new evaluation (analysis). When a new problem, why was your problem crucial? Is the trouble still essential today? Does the problem make a difference tomorrow? If a new criteria or new evaluation (analysis), what are the improvements above previous methods or evaluations? How do they come up with the newest algorithm or perhaps evaluation?
The primary contribution of the paper is that Airavat creates on necessary access control (MAC) and differential personal privacy to ensure untrusted MapReduce computations on delicate data do not leak private data and provide confidentiality, integrity, and privacy warranties.
The aim is to stop malicious computation providers by violating privateness policies a data provider imposes on the info to prevent leaking information about individual items in the data.
The system can be implemented being a modification to MapReduce as well as the Java online machine, and runs together with SELinux
three or more. (15%) Sum it up the (at most) a few key main ideas (each in 1 sentence. )
(1) Initially work to incorporate MAC and differential level of privacy to mapreduce. (2) Suggests a new platform for personal privacy preserving mapreduce computations. (3) Confines untrusted code.
4. (30%) Critique the main contribution
a. Rate the significance of the paper on a scale of a few (breakthrough), 4 (significant contribution), 3 (modest contribution), a couple of (incremental contribution), 1 (no contribution or perhaps negative contribution). Explain your rating within a sentence or maybe more.
This system supplies security and privacy warranties for given away computations on sensitive data at the ends. However , your data still could be leaked in the cloud. Since multiple devices are involved in the computation and malicious employee can dispatched the more advanced data for the outside program, which poises the privacy of the insight data. Even not to this extent, non permanent data is stored in the employees and those info can be fetched even after computation is done.
b. Rate how persuasive the methodology is: just how can the authors justify the perfect solution is approach or evaluation? The actual authors use arguments, examines, experiments, simulations, or a mix of them? The actual claims and conclusions comply with from the fights, analyses or experiments? Will be the assumptions reasonable (at enough time of the research)? Are the assumptions still valid today? Are the experiments smartly designed? Are there distinct experiments that would be more persuasive? Are there additional alternatives the authors really should have considered? (And, of course , is a paper clear of methodological errors. )
Since the author’s stated on page 3 “We aim to prevent malicious computation providers by violating the privacy policy from the data provider(s) by seeping information about person data things. ” Each uses differential privacy mechanism to assure this. One interesting solution to data leakage is that they have mapper designate a range of its important factors. It seems like the fact that larger your details set is definitely, the more level of privacy you have must be user affects less with the output, in the event that removed. That they showed effects that were genuinely close to totally with the added noise, it seems like this is feasible solution to protect the level of privacy of your data input
c. What is the most important limitation from the approach?
As the writers mention, one computation supplier could exhaust system this spending budget on a dataset for all various other computation companies and work with more than the fair share. While there is several estimation of effective variables, there are a large number of parameters that must be set to get Airavat to work effectively. This boosts the probability of misconfigurations or configurations which may severely limit the computations that can be performed on the info.
5. (15%) What lessons should researchers and building contractors take away using this work. What (if any) questions does this work keep open?
The present implementation of Airavat supports both dependable and untrusted Mappers, but Reducers should be trusted and in addition they modified the JVM for making mappers independent (using invocation numbers to recognize current and previous mappers). In addition they modified the reducer to provide differential personal privacy. From the data provider’s perspective they must give several personal privacy parameters like- privacy group and level of privacy budget.
6th. (10%) Recommend your improvement on the same problem.
I have not any suggested improvements.
one particular
