5.4 KiB
5.4 KiB
type:: REVIEWS
tags:: #polystore
year:: 2023
venue:: JSS
full-title:: Extending Hybrid SQL/NoSQL Database by Introducing Statement Optimisation Component
date-start:: 15-06-2023 - 18:09
date-submitted:: 04-07-2023
external-links::
status:: DONE
deadline-submission:: 22-05-2023
file::
- [[Highlights]]
- ((649e88c7-a2a2-48d6-a0ae-b8b62f20c7d8))
- Extended with respect to what?
- ((649e9433-c26d-4d1d-b362-e82c9d89841f))
- Improve what?
- ((649e946d-4898-4ccc-bc4c-ed67d2d18d79))
- What architecture?
- ((649e9bca-6a93-4690-b524-e12377ec92bb))
- The limitations of exitsting approaches concerning performance issues have to be better presented. A comparative table could help with this.
- ((649e9dd2-452c-4ec7-96e8-a47f83d53b72))
- It is necessary to show that optimizations of the different component can be done always in parallel.
- It is necessary to add an explanatory example to show a representative example to discuss all the challenges / innovations the proposed approach supports.
- It is not clear how the optimizations shown in Table 1 are performed. Moreover, what are you optimizing? Execution time? Resources allocations? Memory consumptions? ...
-
- ((649ea3ad-9bab-47b2-b016-1a130602119f))
- At this stage is not clear yet who/what does the optimization. The software architecture of the developed system is necessary to be shown.
-
- ((649ea688-37eb-4efe-a3d1-853c4798294f))
- I agree!!! The point is exactly this one. It is not clear what are you optimizing, and what are the software components implementing the discussed optimization.
- The statement optimization component should be better introduced by showing the architectural context where it is located.
-
- ((649ea7ad-938f-4d56-acc6-777ce8a3b251))
- I would make more evident in the Figures what are the components developed by the authors specifically related to the optmization problem.
- Moreover, it is necessary to show an explanatory query to discuss how the different components are involved while executing it.
- ((64a3d756-ef2f-49b7-9400-2b678a5716db))
- -> Consists of
- ((64a3d97f-362b-49a7-98fd-ebef7b13aa44))
- What's the effect / role of such optimisation rules?
- ((64a3d9ce-f38f-4200-a01e-d99143306478))
- These are not evident to me.
- ((64a3dc16-2f5f-4072-969b-2a52eeed144b))
- Technologies belonging to the same NoSQL or SQL class may have different dialects. How are they supported?
- [[Comments]]
- Summary: The paper presents an approach to optimizing the execution of INSERT, UPDATE and DELETE statements to be executed by hybrid polystores, i.e., database infrastructures that use SQL and NoSQL technologies. The paper is about a relevant and interesting problem. Supporting the creation and management of hybrid polystores is crucial nowadays, and as such, the topic addressed by the authors is very important.
- Comments: Despite the relevance of the topic addressed by the paper, I have several concerns about the article that are related mainly to the following issues:
- Novelty: The paper does not compare with any baselines, and in particular, it neglects relevant related work, including at least the following:
- BigDAWG Polystore (https://bigdawg.mit.edu/)
- Typhon Project (https://www.typhon-project.org/papers), in particular, the work done in the context of such a project on the "Polystore Querying" component (e.g., see [1] [2] [3])
- The authors should refine the related work section and the presentation of the strengths of the proposed approach by comparing/discussing it with respect to such existing technologies.
- Presentation: The presentation of the approach has to be improved to present better the goal of the work employing an explanatory and running example since the early sections of the manuscript. Currently, the paper is too lengthy, and it needs concrete examples that, for instance, can help readers understand the architecture shown in Fig. 2 on concrete statement examples.
- Following my previous point, the reader has to wait for the experiment section to understand that the goal of the statement optimization is to reduce the execution time of the input statements. Even though the term "performance" is mentioned many times early in the paper, there is no explicit reference to execution time as the performance characteristic of interest. Moreover, are other performance characteristics like resource allocation or memory consumption of interest?
- Also, the optimization rules shown in Table 2 should be singularly described by considering explanatory statements so the reader can grasp the rational underpinning of the optimization.
- The paper must discuss how the system has been implemented and what technologies are supported. As the authors know, technologies belonging to the same NoSQL or SQL class may have different dialects. What are those that are currently supported?
- To summarize, even though the paper is about a relevant problem, a significant revision is needed to address presentation issues and clarify the article's novelty.
-
- References
- [1] https://doi.org/10.1109/SANER53432.2022.00082
- [2] https://www.typhon-project.org/_files/ugd/d3bb5c_4349c798cf42409d93a4d588ab1f4162.pdf
- [3] https://www.typhon-project.org/_files/ugd/d3bb5c_fd96be98de6e41b28482cd5992997a61.pdf
-