66 lines
4.2 KiB
Plaintext
66 lines
4.2 KiB
Plaintext
Dear Oliver Kennedy,
|
||
|
||
We are delighted to inform you that your paper PocketData: The Need for TPC-MOBILE has been accepted
|
||
for publication in the 2015 proceedings of TPCTC.
|
||
|
||
Please carefully read the included reviews and incorporate their resolutions
|
||
into your camera ready copy. Please strictly adhere to Springer's format for
|
||
LNCS. Your papers will be checked for formatting and whether comments are
|
||
integrated before they are published.
|
||
|
||
We will send out detailed instructions on how to submit your camera ready
|
||
copy in a separate announcement.
|
||
|
||
Please note that the deadline for early registration is July 8th (2 days
|
||
from now).
|
||
|
||
Congratulations again.
|
||
|
||
We are looking forward to seeing you at TPCTC 2015.
|
||
|
||
Raghu and Meikel
|
||
General Chairs TPCTC
|
||
|
||
|
||
----------------------- REVIEW 1 ---------------------
|
||
PAPER: 10
|
||
TITLE: PocketData: The Need for TPC-MOBILE
|
||
AUTHORS: Oliver Kennedy, Geoffrey Challen, Lukasz Ziarek and Jerry Antony Ajay
|
||
|
||
|
||
----------- REVIEW -----------
|
||
The paper sets the foundation for a new TPC benchmark targeted at embedded database engines for smartphones and tablets.
|
||
|
||
The paper is well presented, clear and extensively supported by empirical data.
|
||
|
||
The well researched evidence presented makes a compelling case for why existing benchmarks do not provide adequate coverage for embedded databases and why a new benchmark is needed.
|
||
|
||
|
||
----------------------- REVIEW 2 ---------------------
|
||
PAPER: 10
|
||
TITLE: PocketData: The Need for TPC-MOBILE
|
||
AUTHORS: Oliver Kennedy, Geoffrey Challen, Lukasz Ziarek and Jerry Antony Ajay
|
||
|
||
|
||
----------- REVIEW -----------
|
||
This paper does an in-depth analysis of SQLite applications on Android mobile phones. They use data gathered from actual phones; in some cases with full access to complete trace data with the phone owners’ consent. As such, the data is a treasure trove of mobile phone performance and workload data.
|
||
The authors analyze the data to group the queries by query types, query complexity, etc., and slice and dice the stats in multiple ways to detect patterns and tendencies.
|
||
All in all, an excellent paper! One area the authors might want to focus on is how one can capture the profile of the workload with a manageable number of tables/queries without losing the effects of the long tail. This is a big, common problem for benchmark developers, and with the wealth of data they have, they might be able to tackle it. A second question is whether they have collected enough data to be able to reproduce the profile with a synthetic workload. Quite often people collect table statistics and the query source code, but when a synthetic workload is put together, it doesn’t have the same profile. There are lost subtleties, e.g., the distribution of the numerical values that are inserted in embedded parameters at run time, that may be lost with a static profile
|
||
|
||
Typos:
|
||
“operation” in the abstract
|
||
|
||
|
||
----------------------- REVIEW 3 ---------------------
|
||
PAPER: 10
|
||
TITLE: PocketData: The Need for TPC-MOBILE
|
||
AUTHORS: Oliver Kennedy, Geoffrey Challen, Lukasz Ziarek and Jerry Antony Ajay
|
||
|
||
|
||
----------- REVIEW -----------
|
||
This paper presents an experimental evaluation on the need for evaluating smartphone embedded databases. The paper focuses on Android and SQLite, a combination widely used in the smartphones market. In practice, the paper presents the results from a one-month trace of SQLite activity on 11 smartphones. The motivation for this work is two fold: 1) queries and access patterns in Android SQLite are quite different from canonical server workloads: 2) the performance of the embedded database impacts the performance of the system as the database is accessed by the OS and also by other installed apps.
|
||
|
||
The paper starts by contextualizing and motivation the study. Afterwards, the experimental setup is introduced and the complexity of the queries observed during the study is discussed. The paper continues looking at the overall runtime characteristics of the workload observed. Related work is presented before an interesting discussion on the need for a TPC benchmark focusing on mobile platforms grounded on the observed results.
|
||
|
||
This is a quite interesting paper. It is well written and easy to understand. The paper achieves the intended goal of motivating TPC to work on a mobile benchmark.
|