Ru-containing polyoxometalate fabricated on graphene oxide: Preparation, characterization and catalytic activity

Article abstract of DOI:10.1016/j.jssc.2019.03.043

Polyoxometalates (POMs) with remarkable redox characteristics and good stability are considered as versatile catalysts, meanwhile graphene oxide (GO) with a layered structure and myriad oxygen atoms on its surface is also good catalyst. Thus their combination will have the synergistically enhanced effect on the catalytic performance and the reusability of the heterogeneous catalysts. Here, an environmentally-friendly heterogeneous catalyst (GO/PEI/RuPW10) was prepared by fabricating [(γ-PW₁₀O₃₆)₂Ru₄O₅(OH)(H₂O)₄]^9- (RuPW10) on the surface of GO functionalized by polyethylenimine (PEI). Further, this catalyst was applied to n-tetradecane oxidation using air as the oxidant. The results showed that 4.73% RuPW10 loaded catalyst exhibits high conversion (50.45%) of n-tetradecane under the optimized reaction condition. A recyclability test suggested that this catalyst can be reused up to five times without significant changes for the conversion of n-tetradecane. As a result, due to the high conversion, better stability and reusability, immobilizing POMs onto the surface of GO modified by PEI is one of the effective means to develop heterogeneous catalysts.

Full text of DOI:10.1016/j.jssc.2019.03.043

Evaluating Distributed Transactional Database System
Pengwei Ma
Kai Wei
Chunyu Jiang
China Academy of Information and
Communications Technology
52 garden north road
Haidian District, Beijing, China
0086-18500188116
China Academy of Information and
Communications Technology
China Academy of Information and
Communications Technology
52 garden north road
5
2 garden north road
Haidian District, Beijing, China
086-15201365784
Haidian District, Beijing, China
0086-18600832598
0
mapengwei@caict.ac.cn
weikai@caict.ac.cn
Jiangchunyu1@caict.ac.cn
correctness and performance. According to statistics, Taobao's
double Eleven" turnover increased 3000 times in eight years, that
is, from 52 million yuan in 2009 to 168.2 billion yuan in 2017.
However, in the 2011-2014 "double Eleven" process, there were
congestion in 8 of the 16 listed banks. The traditional database get
limitation to support new business scenario like mobile payment.
ABSTRACT
"
With the rapid development of mobile Internet and electronic
commerce, massive transactional operations under new data
modules becomes the dominated workload of the database
systems. In order to meet the functional and performance
requirements of this kind of business, new distributed
transactional databases system evolve fast. However, because of
the distinct technical differences between the new distributed
transactional database and traditional centralized transactional
database, such as technical component architecture and principles
of computing and storage, the evaluation framework and method
for traditional database cannot directly apply to distributed
transactional database. For example, the existing benchmarks
can’t evaluate high availability, scalability, and distributed
consistency issues which are important characteristics after
distribution. It is urgent for the industry to construct evaluation
methods for distributed transactional databases so as to
comprehensively measure the technical maturity and capability of
these products. In this paper, we first summarize the definition
and characteristics of distributed transactional database system.
After evaluating the existing test benchmarks for distributed
transactional database, we propose a new evaluation framework
and test methods for distributed transactional database system.
This paper provides a new guidance for testing and selecting
distributed transactional database in the world.
At present, large enterprises, such as financial banks, mostly adopt
centralized transactional database architecture. This architecture
adopts scale-up mode, that is, increasing the hardware
performance of a single machine to increase the upper limit of
processing power. However, with the gradual failure of Moore's
law, the scale-up scheme began to reveal performance bottlenecks.
Due to cost and performance reasons, most of the emerging
Internet enterprises began to explore the road of distributed
transactional database architecture. And through many years of
practical experience, it is proved that the distributed transactional
database architecture can support the large-scale and high
concurrency business scenario [2]. Taking Alipay as an example,
the whole line of IBM minicomputers and EMC storage devices
were removed in 2013, and a new distributed database was
developed to replace Oracle databases in 2014. Oceanbase and
x86 servers architecture successfully passed the 2015-2017 three
years of online business test. The corresponding hardware and
software costs are reported to have fallen by about 60%, while
performance has improved dozens of times. Therefore, the
distributed transactional database will be the evolution direction
of the transactional database in the new business environment.
CCS Concepts
•
Software and its engineering ~ Formal software verification
However, the development of distributed transactional database
technology is still at the beginning stage, so there are still many
problems [3]. The main problems are listed below: 1) there was
not a unanimous definition about distributed transactional
database. Each company develop their own product based on their
Keywords
distributed transactional database, evaluating, financial.
1
. INTRODUCTION
requirements and understandings.
2) the development of
The time of electronic commerce is coming. I Internet finance and
mobile Internet make payment habit change from cash to mobile
payment, which results in a large number of live users, data and
concurrent volume of business scenarios [1]. These business
scenarios require transactional databases to support their
distributed transactional database technology is still isolated,
which means that the related achievements and experiences of
industry have not been efficiently shared. 3) distributed
transactional database gets new features, the industry call for new
benchmark tools to measure all kinds of technical realization.
Permission to make digital or hard copies of all or part of this work for
personal or classroom use is granted without fee provided that copies are
not made or distributed for profit or commercial advantage and that
copies bear this notice and the full citation on the first page. Copyrights
for components of this work owned by others than ACM must be
honored. Abstracting with credit is permitted. To copy otherwise, or
republish, to post on servers or to redistribute to lists, requires prior
specific permission and/or a fee.
In order to solve the above problems, this paper firstly studies the
technology and industry of distributed transactional database, and
gives the definition of distributed transactional database. At the
same time,
a comparative study of various distributed
transactional database technology routes is carried out, and the
technical characteristics are compared and summarized. Then, the
shortcomings of the current transactional database testing
benchmark are analyzed. and we developed the functional and
performance evaluation methods of distributed transactional
database which covers all kinds of technical features. This paper
ICCMS 2019, January 16–19, 2019, Melbourne, VIC, Australia
ACM ISBN 978-1-4503-6619-9/19/01…$15.00
DOI: https://doi.org/10.1145/3307363.3307364
2
03

makes some suggestions for the development of distributed
transactional database.
to be able to support high concurrent performance requirements,
so as to truly meet the requirements of the business development.
d) Support for standard SQL (standard query language) data
manipulation
2
. OUTLINE
The rest of this paper is structured as follows: In Sect.3 we
introduce the standard definition and typical technical
characteristics of distributed transactional database. Section 4
analyzes the advantages and disadvantages of the current
distributed transactional database technology architectures. Next,
Sect.5 analyzes the shortcomings of the current transactional
database evaluation methods in the current environment. Sect.6
gives the implementation principles of the distributed
transactional database testing benchmark and introduces the
technical specifications of our functional evaluation methods for
distributed transactional database. Finally, we describe how the
evaluation methods of distributed transactional database will be
used in the future in Sect.7.
At present, most of the database services are written in the SQL
language. If the distributed transactional database can’t support
complete standard SQL, then the stock business will face large-
scale transformation, which is intolerable to the database
application. Therefore, the ability of supporting standard SQL in
distributed transactional database is the precondition of database
distributed transformation.
e) High availability guarantee for automatic Fault Self-Healing
and switching
After the transactional database is distributed, the pressure of
computing and storage is shared by each node. At this time, if a
single node fails, the database system needs to perform the
switching process quickly to ensure the continuity of business.
3
. DEFINITION OF DISTRIBUTED
f) Support scale-out ability across data center levels
TRANSACTIONAL DATABASE
At present, there is no uniform definition of distributed
transactional database in the industry. This paper summarizes the
distributed transactional database into three core attributes and six
core capabilities after summarizing the views of academia and
industry.
Due to the continuous growth of transactional business scale, the
distributed transactional database will be put forward higher and
higher performance requirements. Because distributed mainly
depends on scale-out ability to improve processing performance,
distributed transactional database needs horizontal expansion
ability at the level of data center to meet the needs of future
business growth of transaction processing.
3
.1 Three core attributes
The three core attributes of a distributed transactional database are
as follows:
4
. ANALYSIS OF CURRENT
a) Transactional database properties
DISTRIBUTED TRANSACTIONAL
The database core data logical management model has the
transaction ability, conforms to the complete ACID (atomicity,
consistency, isolation, persistence) attribute request.
DATABASE
This chapter will analyze the current development status of
distributed transactional database from two aspects of technology
and industry.
b) Massive Online transaction processing capability
A concerned database management system that can handle
massive online transaction processing loads.
Since the distributed transactional database is still in the early
stage of development, and each database product is gradually
developed based on its own business practice, different technical
routes and theoretical frameworks have been formed. Database
theorists have put forward a new data management model, trying
to fundamentally solve related problems. However, the database
industry tries to find a breakthrough in the database design
architecture through the practice of various engineering research
and development methods [4].
c) Distributed architecture attribute
Both computing and storage make effective use of the parallel
capabilities through multiple nodes.
3
.2 Six core capabilities
The six core capabilities of a distributed transactional database are
listed as follows:
Combining the academic and industrial achievements in the field
of distributed transactional databases in the past 10 years, we can
sort out the development direction of several kinds of distributed
transactional databases.
a) Computing and storage can be implemented in hardware based
on industrial standards such as X86
The new distributed transactional databases are based on the
standard X86 industry standard hardware, but not on the original
mainframe and minicomputer proprietary hardware. This is
helpful to control the budget and unify the development method,
so that it can be widely popularized in the industry.
Table 1. Table captions should be placed above the table
Primary
Secondary
Tertiary
Technical
characteristics
classificat classificatio classificat
ion
n
ion
b) Comply with the relational model of the transactional database
and fully meet the ACID requirements of the transaction
The business layer
intrudes strongly
into the business
layer in order to
realize the
Distributed transactional databases are mainly used to support
transactions. Distributed transactional databases need to be able to
fully support the ACID attribute.
Based on
centralize
d
Depending on the
application end to
complete the
c) Ability to support highly concurrent OLTP workloads
transactio
nal
transformation
distributed
correlation
With the rapid development of mobile Internet, the number of
OLTP services supported by distributed transactional databases
will increase exponentially. The distributed transactional database
needs not only to ensure the correctness of the business, but also
database
characteristics.
Limited intrusion
into the business
Middleware Implemen
based
tation of
2
04

transformati distributed
layer
c) New distributed Database Theory and Engineering
implementation Architecture
on
characteri
stics in the
Intermedi
ate layer
embedded
in the
This kind of database is usually composed of four parts: database-
driven, computing node cluster, data node cluster and global
management node [6]. Multiple copies of the data are stored on
multiple nodes of the data cluster in a customized or automatically
optimized strategy. The data nodes do not share the data, and the
synchronization between the replicas is realized by the automatic
consistency algorithm. The computing node is responsible for
distributed query processing and global transaction control. The
global management node is a necessary component to realize the
consistency of distributed transactions and to maintain the
partition and topology of the sub-table. The nodes communicate
with each other through a high-speed interconnected network,
thus completing the rapid processing and response to the
application data requests.
Applicatio
n layer
Implemen
tation of
distributed
feature
based on
Independe
nt
Middlewa
re
Limited intrusion
into the business
layer
5
. ANALYSIS OF CURRENT
New Computational
Storage algorithm
and Engineering
implementation of
distributed
correlation, weak
intrusion to the
Service layer
TRANSACTIONAL DATABASE
BENCHMARKS
Transactional database has accumulated a lot of test experience in
the long process of development [7]. Many organizations and
individuals rely on this testing experience to create common
testing tools and benchmarks to verify the functionality and
performance of transactional databases. First of all, this chapter
briefly introduces the transactional database testing tools which
are commonly used in the industry[8].
Based on New Database Theory
The detailed descriptions of three distributed transactional
database technology solutions are as follows:
a) Business distributed transaction Transformation Scheme based
on single transaction Relational Database
a) Sysbench.
Sysbench is an open source multithreaded performance testing
tool that performs performance tests in areas such as CPU /
memory / thread / IO / database. The database currently supports
MySQL / Oracle / PostgreSQL. It is a very popular dba pressure
testing tool.
This scheme can effectively utilize the maturity advantage of the
original single computer transaction processing database. By
establishing the rules of data sharding and data routing in the
application layer, a composite distributed transactional database
architecture is established [5].
However, it lacks scenario support and can only be used to test
common resource performance limits.
This architecture requires the DBA (Database Administrator) to
manually control the distribution and management of application
traffic. Therefore, the intrusion into the business layer is more
serious, but it can alleviate the disadvantage of the lack of storage
and computing power of centralized database.
b) TPCC
TPC-C was developed by the TPC Institute to test the
performance of a typical complex OLTP system. TPC consists of
five processing logics. It is a pressure measuring tool which is
close to the business of e-commerce platform. It is the most
commonly used benchmark tool in testing transactional database
at present. Its execution logic includes New-Order, payment,
Order-Status, delivery, Stock-Level.
b) Solution of distributed Management component based on single
transactional database
By improving the original architecture of single transactional
database engine layer, the independent distributed control logic
unit is used to implement the automatic routing strategy. The
consistency problem of distributed transactional database is
solved by matching theory and algorithm of consistency. This
method is mainly divided into embedded middleware mode and
independent middleware mode two kinds of solutions.
However, TPCC can not fully reflect the ability of distributed
transactional database because of its slow data generation, single
business scenario, and lack of distributed characteristics to
measure dimensions and so on.
c) Mysqlslap
Embedded middleware uses DAL (Data Access layer) to manage
data fragmentation rules and routing strategies. The DAL here is
not a standalone middleware. It is a middle-tier pattern embedded
in the business side and is typically provided to the application
system for invocation in the form of a Jar package.
Mysqlslap is an official stress testing tool that has been available
since version 5. 1. 4. Stress tests are performed by impersonating
multiple concurrent clients accessing MySQL, while providing a
more detailed data performance report. It can be used to compare
the concurrent pressure performance of multiple storage engines
in the same environment.
Independent middleware manages data fragmentation rules and
routing policies by proxy server mechanism. Independent
middleware is usually a separate logical and physical processing
system architecture. It is isolated from the business layer and the
data storage and computing layer, focusing on the data
distribution management function.
However, mysqlslap still lacks distributed features to measure
dimensions, and the scenario is single. At the same time, it is not
compatible with non-database of mysql class.
d) tcpcopy
2
05

TCP copy is a real-time replication tool for tcp traffic, version 1. 0
developed and maintained by NetEase engineer @ tcpcopy. It is
generally used to copy the online traffic of the production
environment to the test environment in real time for testing. The
advantage is that the test data is close to the real level and
relatively simple to implement.
b) implicit conversion of data types.
c) Total migration of heterogeneous data in traditional databases.
d) Ability to use other databases for remote connection operations.
e) The supporting ability of Common hardware such as x86.
6
.3 Management
However, it is complex to configure, and involves the online
environment, so the risk is high.
Complete and friendly management ability can provide effective
support for the operation and maintenance of distributed
transactional database. This helps database DBA managers,
database business leaders, and related application support teams to
have better database management and control capabilities.
Generally speaking, current testing schemes and tools commonly
used in industry cannot be applied directly to distributed
transactional databases. Based on the advantages of the current
common test tools, the industry needs to supplement the
distributed characteristics and develop functional and performance
evaluation methods and tools suitable for distributed transactional
database.
Specific indicators of management capacity include:
a) Easy-to-use installation deployment, configuration management,
and upgrade capabilities.
b) Distributed transactional database is capable of real-time
monitoring and statistical analysis of nodes and data. At the same
time, distributed transactional database can notify and alarm all
kinds of information in a variety of forms.
6
. FUNCTIONAL EVALUATION
METHODS OF DISTRIBUTED
TRANSACTIONAL DATABASE
The computing and storage mode of distributed transactional
database is different from that of traditional centralized
transactional database [9]. Therefore, the traditional centralized
transactional database evaluation methods can’t be directly
applied to the distributed transactional database.
c) Distributed transactional database can realize no deadlock or
automatic processing of deadlock.
d) Distributed transactional databases can manage, view, and
process common types of logs.
e) Distributed transactional databases can support online database
schema definition languages and management commands.
According to the characteristics of distributed transactional
database, this paper puts forward the functional test method of
distributed transactional database from six dimensions, so as to
measure the capability of distributed transactional database in all
directions. It provides guidance for the development and
application of distributed transactional database.
f) Distributed transactional databases are capable of managing,
viewing, and processing common types of logs in the system.
g) Distributed transactional database can backup and restore data
in many forms, such as timing, increment and full quantity.
The evaluation methods in detail is as follows.
h) Distributed transactional database can customize the isolated
grouping of resources within the cluster.
6
.1 Basic Function
Distributed transactional database should be able to have the basic
functions of traditional centralized transactional database [10].
Thus, after the distributed transformation, the adaptation work of
the business system and the personnel can be simplified to the
maximum extent, and the normal operation of the existing
business of the database can be guaranteed. At the same time,
distributed transactional database needs to be able to reflect the
basic distributed capabilities. This prepares the technology for
performance extensions and new architectures such as operations
and maintenance.
6
.4 High Availability
Distributed transactional database is the infrastructure of a variety
of core businesses, which may lead to a large number of economic
property losses when failure occurs. Therefore, the high
availability must be able to meet the corresponding requirements,
thus ensuring the continuity and stability of the business.
Specific indicators of high availability include:
a) Distributed transactional database can provide normal database
service in hardware accidents such as power outage, hard disk
failure, network flashover, computer room disconnection and so
on.
Basic functional indicators include:
a) Support capabilities for common data types, operators,
character sets, functions, SQL syntax, etc. The ability of data
slicing, partitioning, distributed transaction, isolation level setting.
b) Support ability of data fragmentation, partition operation,
distributed transaction, isolation level setting.
b) Distributed transactional database can provide normal database
service under the failure of primary and secondary center data
backup and synchronization.
c) Distributed transactional database can provide normal database
service under CPU resource overload, IO resource overload,
memory resource overload, disk space overload and other
operating system software failures.
c) Ability to execute plans, table partitions, and indexes.
6
.2 Compatibility
Compatibility represents the ability of a distributed transactional
database to connect with existing business systems and common
database hardware and software tools [11].
d) Distributed transactional database can provide normal database
service under database service failure such as data file corruption,
data node corruption, system table corruption and so on.
This class of indicators measures the scale of impact on the
existing business ecosystem after the implementation of a
distributed transactional database.
e) Distributed transactional database can provide normal database
service in the case of application connection loss, peripheral
system error and so on.
Compatibility indicators mainly include:
a) Support ODBC, JDBC and other common connections.
2
06

distributed transactional database. In the future, our team will
automate the evaluation methods in the form of tools, and provide
more convenient services for the selection and evaluation of
distributed transactional databases.
6
.5 Scalability
Distributed transactional databases must provide online horizontal
scaling capabilities [12]. It needs to meet the needs of customer
business growth on the basis of customer business expansion
without obvious awareness, so as to solve the bottleneck in disk
capacity and computing performance of the existing centralized
transactional database system.
8
. REFERENCES
[
1] Corbett J C, Dean J, Epstein M, et al. Spanner: Google's
globally-distributed database[C]// Usenix Conference on
Operating Systems Design and Implementation. 2012:251-
Specific references to capacity expansion include
2
64.
a) Computing resources for read and write services can be
automatically balanced across nodes in a cluster.
[
2] Cheung D W, Ng V T, Fu A W, et al. Efficient mining of
association rules in distributed databases[J]. Knowledge &
Data Engineering IEEE Transactions on, 1996, 8(6):911-922.
b) The stored data can be balanced according to the customized
requirements of the business side.
[
[
[
3] Zsu M T, Valduriez P. Principles of distributed database
c) The cluster can realize online expansion operation without
affecting the normal operation of the business.
systems[J]. Springer Berlin, 1999, 24(1):31-41.
4] Bernstein P A. Concurrency Control in Distributed Database
Systems[J]. Acm Computing Surveys, 1989, 13(2):185-221.
d) Cluster can be configured to achieve online capacity reduction,
so as to achieve reasonable planning and utilization of resources.
5] Romano P, Romano P. A uto P lacer : Scalable Self-Tuning
Data Placement in Distributed Key-Value Stores[J]. Acm
Transactions on Autonomous & Adaptive Systems, 2014,
6
.6 Safety
In order to keep corporate business away from legal and business
risks, distributed transactional databases need complete security
capabilities. For example, to prevent sensitive data disclosure and
eliminate the wrong operation behavior and so on.
9
(4):19.
[6] Kumar R, Gupta N, Charu S, et al. Manage Big Data through
NewSQL[C]// National Conference on Innovation in
Wireless Communication and NETWORKING Technology.
2014.
Specific indicators of security capacity include:
a) Distributed transactional databases are able to verify
permissions for data operations at the library level and at the table
level.
[
7] Kumar R, Parashar B B, Gupta S, et al. Apache Hadoop,
NoSQL and NewSQL Solutions of Big Data[C]//
International Journal of Advance Foundation and Research in
Science & Engineering. 2014.
b) Distributed transactional database can authenticate access users.
Such as implementing whitelist and blacklist reviews.
c) Distributed transactional database can audit and trace all kinds
of operations performed, so as to locate and analyze the key
problems.
[8] Ardekani M S, Terry D B. A self-configurable geo-replicated
cloud storage system[C]// Usenix Conference on Operating
Systems Design and Implementation. USENIX Association,
2
015:367-381.
d) Distributed transactional database can be used for flow control,
so as to prevent the whole system failure caused by a certain burst
of traffic which will affect the normal operation of other services.
[9] Ling G U, Weng H X, Hua-Liang H U, et al. OceanBase
schema design for OLAP application[J]. Journal of East
China Normal University, 2014.
7
. CONCLUSION
[
10] Ramanathan S, Goel S, Alagumalai S. Comparison of Cloud
database: Amazon's SimpleDB and Google's Bigtable[C]//
International Conference on Recent Trends in Information
Systems. IEEE, 2011:165-168.
In this paper, the distributed transactional database is studied
deeply. Firstly, this paper proposes a comprehensive definition
and feature definition of distributed transactional database.
Secondly, this paper analyzes the current distributed transactional
database of all kinds of technical architecture and product
characteristics. Based on the analysis of the shortcomings of the
distributed transactional database evaluation method, this paper
proposes the distributed transactional database function and
performance evaluation method. Based on the distributed feature,
the method measures the special ability of distributed
transactional database from multi-dimension. These works have
very important guiding significance for the development of
[11] Sukhwani B, Min H, Thoennes M, et al. Database analytics
acceleration using FPGAs[C]// International Conference on
Parallel Architectures and Compilation Techniques. IEEE,
2017:411-420.
[
12] Abbas Q, Shafiq H, Ahmad I, et al. Concurrency control in
distributed database system[C]// International Conference on
Computer Communication and Informatics. IEEE, 2016:1-4.
2
07