Valery Sadovoy, Assistant Professor

Crimea State Medical S. I. Georgievsky University, Ukraine

Oleg Parmenov, Assistant Professor

Taurida National V.I.Vernadsky University, Ukraine

Ibrahim Ismailov, Master of Computer Science

Taurida National V.I.Vernadsky University, Ukraine

 

Distributed information system

“Register of myocardial infarction”

 

The task was to design distributed information system “Register of myocardial infarction”. This system is very crucial, because cardiovascular diseases are common reason of death today [1]. In order to provide data processing and statistical analysis, it was necessary to automate the process of data maintenance. The main customer demands were: a) to provide multi-user access to a data repository from different regions of the Republic of Crimea;  b) to provide data processing and data analysis functionality.

 As usual, the property named “distributed” means that the system operates with a regardless of a user’s location [2], that`s a distributed database is a system in which all components are located on networked computers. These computers communicate with another one and coordinate their actions by passing messages. Another widely used conception of database management is based on linked network nodes.

There were the following requirements to the distributed computer system:

1)       To provide multiuser access via the Internet, i.e. the system must be based on properties of non-concentrated database implementation with Web interaction. An application user doesn't interact right to the server to provide transitory offline service (without an access to the Internet) with synchronization in a background;

2)       The system has to support statistical and regression analysis of stored data.

 

Thus it was important to solve three technical problems:

1)       to choose web hosting service. We take DigitalOcean service because of its inexpensive and reliable hosting. Minimal delivered requirements (512 MB operational memory, 2 cores) are quite acceptable at that period of time.

2)       to choose database management system. The preference was given to MySQL database system because of its performance and gratuity.

3)       to choose software environment. The Java has an advantage because of its gratuity.

We propose the synchronization technology between a local database and server data repository which is grounded on the monitoring of the database version. This monitoring is implemented on a local nodes and on the server.

All local copies of database generate unique identification number– shortly called ID or version number. These ID are stored on every node after any local database modification. The similar technology of ID generation is implemented in the server database.

When a single user adds or modifies data in a local database, the application sends the request to the server for updating automatically. Procedure of updating server repository works in the same way.

On account of synchronization and personal data protection requirements, we embed the following software and hardware solutions:

1)   Data encryption. All data is stored and is sent via Internet in ciphered form with 8-byte DES algorithm encryption;

2)   Data access security. The data safety is based on principles of enclosed system. Firstly, data can be obtained and correctly decrypted only by using this software and not in the other way. Secondly, the password is need for installing and executing the application on a local computer.

We estimated a functional stability, survivability and reliability of the system. In accordance to the requirements of standard ISO/IEC 27001, we carried out comprehensive testing to ensure reliability such as: software regression testing and performance optimization, load testing and Denial-of-Service attacks modelling (DOS-attacks may exhaust the server resources and causes a loss of operability and system crash).

By tuning-up the database configuration and hosting service configuration we get steady stability of application. System stability and data integrity are provided even in over-load system mode. Testing procedure loads the server component by a simulation of about 100 virtual users activity.

It has been found that in load peaks with high-intense transactions the system performance does not violate. The characteristics of system reaction and system load which obtained in test mode are shown in the graph.

In this test, the interval between consecutive transactions of an every of 100 users was about a second. Obviously, the system response time increases with the number of users, but it’s tending to stabilize at about 50 users. Even with 100 users loading the response time does not exceed 5 seconds.

The system provides a highly efficient statistical analysis using a variety of common criteria. We implement prevalent mathematical methods to provide calculating of Student data reliability criteria, Spearman’s and Kendall’s correlation ranks and a regression analysis. The diagram presents the regression analysis results as an example of data processing.  Thus the clotting time decreasing has a direct correlation with high rates of atherogenic (left chart) .

 

The diagram of risk factor categories - heart rate and cholesterol level - is shown on the right chart. User application provides two-dimensional and three-dimensional charts and diagrams with an approximation on squares method and non-linear multivariable regression which is solved by normal equation.

 

Main gained results are:

1)       It was designed the system for cardiology hospitals of the Crimea which works in real-time with Internet access to database repository; it provides an offline-mode operating also.

2)       It was obtained database and server stability and their high performance. The system guarantees from unauthorized access.

3)       The software suite also provides data processing and analyzing. An application user may select on one’s own criteria the method of data visualization and statistical processing procedure.

We hope that the system increases the efficiency of diagnostic work and reduces a medical errors chance.

 

References

[1]. Roger VL, Go AS, Lloyd-Jones DM, et al. (January 2012). "Executive summary: heart disease and stroke statistics--2012 update: a report from the American Heart Association". Circulation 125 (1): 188–97

 [2]. M. T. Özsu and P. Valduriez, Principles of Distributed Databases (3rd edition) (2011), Springer