DE102011056595B4 - Method for supporting request-based server programs - Google Patents

Abstract

A method for supporting request-based server programs in a client-server software system, wherein a module is present on a server, in particular a server software and the module is called when requesting a user to a request-based server program, and in turn from the Request extracted the identification of the client environment to which the request is directed, the module using a mapping table the request based on the tenant identification assigns a lot of configuration parameters, where it adjusts the configuration of the request-based server program and reloads, so the request for the data areas provided to the user is accessed within the external processes for status and persistence data, the software module handing over control to the server program after the connection has been established, which then starts processing the request as would have the reconfiguration n never found.

Description

The invention relates to a method for supporting request-based server programs in a client-server software system, a computer program product and a computer system.

In the prior art request-based server programs are described. The request-based server programs are computer programs that are installed on server hardware and are constantly running there. They accept requests from users and process them using the data and resources available to them that are available through hardware or connected systems. Inquiries are often made using the Hypertext Transfer Protocol (HTTP). As a client program, the user uses a web browser. Requests are generated by directly typing in and calling an HTTP address or by clicking on a link to an HTTP address (so-called link).

As a rule, request-based server programs have status and persistence data that provide users with the added value of a full-fledged software program without having to install it on the user's client machine. Status and persistence data are kept in external systems such as database management systems or key-value storage. Communication with these systems takes place via well-defined channels (for example via Unix sockets or TCP / IP) between the latter and the request-based server programs. For the user of the request-based server programs, the only requirement is the installed web browser, which is preinstalled as a universal tool anyway on most computers.

Regular operators of request-based server programs are companies whose employees use the software jointly for communication and collaboration, for example. For such companies, it is often not useful from an economic point of view, to purchase server hardware and connect to the Internet, since even a single server computer often oversized and the purchase or rent in whole is not economical, since the resources of the server computer lay fallow.

The operator uses the service of a provider ("provider"), in which the request-based software is installed on a server hardware that is also used by other clients of the provider in order to reduce costs. In this context, the client is the unit that the operating company itself and its employees form together.

For parallel execution of the various request-based server programs of its clients on a physical server hardware for the provider according to the current state of the art, two approaches in question. Purpose: By optimally deploying and utilizing available server hardware, server programs for different clients should be provided so that each client can fully utilize its own instance of software and neither (a) interfere with the use of other clients nor (b) gain access to existing data of other clients.

The first approach describes virtualization at the operating system level through hypervisor ("Virtual Private Server"). Through the use of a hypervisor for each client own complete operating system environments are created. The resulting guest environments run within the host environment as their own processes and form the client to a self-contained environment that is comparable to that of a full-fledged physical server computer.

The client is free to install and use server programs within the guest environment and has full administrator access to their guest environment to use operating system services, etc. The client has no ability to access the guest environments of other tenants within the same host environment.

This type of deployment is suitable for clients with very individual requirements. These can be implemented individually, since the environments of the individual clients do not share resources.

Maintaining these environments - a task that is often done by the provider - is very time-consuming for this because there is little commonality between tenant environments and each environment needs to be maintained individually.

The disadvantage of the approach is that the resource consumption (memory, persistent storage, computing power) is very high, because in addition to the resources for the server programs for each client resources for the execution of the guest operating system and its processes must be provided.

As a result, only a small number of clients per physical server machine can be accommodated. The number of tenants per hardware is limited by available memory and computing power. This leads to comparatively high costs per client on the part of the provider.

Another approach is the parallel installation within the same operating environment ("shared hosting"). In this case, within the environment which is installed by the one operating system which runs on the server computer, the various instances of the server program are installed in different directories in the file system and are started from there to provide services to each client for accessing clients. In some cases, chroot environments are also used here, which serve to isolate the individual clients from each other. However, chroot environments have no positive effect on resource consumption.

The client can install and use server programs within its scope in the file system, but is restricted in its freedom because it does not have full administrator rights. Certain operating system services, programming languages and programs (such as compilers) are not available to him, as their use could lead to other clients being affected or data being accessed without permission. The client may have the ability to access the other tenants' panes, for example if they have improperly set access permissions to files and directories due to carelessness.

This type of deployment is suitable for clients with individual requirements that are feasible within the limitations of operating system services, programming languages, etc.

As with virtualization (first approach), maintaining these environments is very time-consuming for the provider because there is little commonality between tenant environments and each environment needs to be maintained individually. At least, however, you do not have to maintain your own operating system instance for each client.

Again, it is disadvantageous that the resource consumption (memory, persistent storage, processing power) is still quite high, since the different instances of the server programs that have been installed for each client, each block their own resources.

The number of tenants that can be accommodated per physical server machine is higher than in virtualization, but still limited by available memory and computing power. It can be observed that this variant is frequently chosen when the various clients have very similar requirements, such as the use of the same request-based programs.

The US 2002/0120724 A1 discloses a mechanism for reconfiguring a server. It uses a pointer-based technique that allows server programs to reload their own configuration without having to reboot. However, it is not possible to wait until execution of a request until the reconfiguration is complete, then to process only this one request with the new configuration and then reconfigure the next request.

Also the US 2003/0101245 A1 discloses a dynamic reconfiguration of a server without these capabilities.

US 6,343,360 B1 discloses a method of configuring computer systems based on mapping tables using, for example, postal codes as key-values. This document does not describe that the reconfiguration of the connection to the persistence layer can be used with every request, in principle not necessarily using multi-tenant server programs for multi-client operation.

US 2006/0123121 A1 discloses a session information management system to ensure consistency of session information between different servers. Also, this document does not disclose any teaching to eliminate the disadvantages of known methods for using multi-mandate query-based server programs.

The invention therefore has the task of creating a server environment within which the available hardware resources are optimally utilized in the event that different clients have very uniform requirements for the respective request-based server programs. As a precondition, as described above, the request-based server programs for holding status and persistence data are based on external systems such as, for example, database management systems or key-value storage. As an example, here is the case in which one and the same request-based server program for different clients to be provided on a server hardware. Nevertheless, the individual clients in the sense of (a) and (b) (see above page 2 ) be isolated from each other.

The problem is solved by the independent claims. Advantageous embodiments can be found in the subclaims.

It has been quite surprising that a method of supporting request-based server programs in a client-server software system may be provided that does not have the disadvantages or deficiencies of the prior art. In the process, a special software module is installed on a server. This module is invoked every time a user requests a request-based server program and, in turn, extracts from the request the identification of the tenant environment to which the request is directed. This identification is in particular a host name. Then, the module assigns a set of configuration parameters based on the tenant identification with the help of an allocation table of the request, with the help of which it then adjusts the configuration of the request-based server program and reloads, so that for the request to the ready for the user data areas within the external processes for status and persistence data. After establishing the connection, the module passes control to the request-based server program, which then starts processing the request as if the reconfiguration never took place.

It was completely surprising that the resource consumption of the server, in particular the server hardware per client can be reduced by the inventive method. A person skilled in the art is aware of a server which, within the meaning of the invention, may also be referred to as a "host". The server preferably includes a computer that runs essentially permanently. Therefore, it is preferred to use components designed for continuous operation, for example SAS hard disks with a RAID system, several main processors and redundant power supplies. In addition, the server must have high resources (memory, persistent storage, processing power), so that parallel processes are fast executable.

In particular, the limited resources such as main memory, persistent memory and computing power can be saved by the method according to the invention, so that fewer resources have to be expended for a request or else for parallel requests. That is, the process has a technical effect on the technical elements of the server - it essentially leads to resource savings, so that the performance but also the durability of the resources and inevitably the server can be increased in an extreme way. This represents a technical advantage that has not been achieved by the prior art.

For the purposes of the invention, a module refers to a closed functional unit of software consisting of a sequence of processing steps and data structures. Content of a module is preferably a recurring calculation or processing of data that must be performed multiple times. It is preferred that isolation of interface and implementation is achieved by means of the module, the interface of a module in particular defining the data elements needed as input and result of the processing by the module and the implementation preferably containing the actual program code. The preferred module in the sense of the invention may be referred to in particular as a function or subroutine which is called by the server program and performs a series of processing steps and supplies data as a result to the calling server program. It may also be preferred that the module calls itself further modules, so that a hierarchy of program calls is possible.

It was completely surprising that the module which describes the invention makes it possible to reuse a process which is already in the working memory, that is to say a running process, without the code having to be stored redundantly in the file system and loaded into the main memory. The inefficient use of the available persistent memory and main memory in the prior art can thus be avoided: while in the prior art one and the same server program at least once per client in the file system of persistent storage and at least once per client loaded into the main memory had, it occupies only once in the system described by the invention persistent memory and is also loaded only once in the main memory. The underlying method of the module can be implemented and compiled in many programming languages and provided in the form of program libraries, so that it can be used universally and installed on different systems. Thus, it can be used, for example, for Linux, Mac or Windows based server programs and environments.

In addition, the module can also be retrofitted into existing server programs without significantly affecting the functionality of the server programs.

It was further surprising that many programming languages embody the concept of the module described by the invention support integrated language means such as C, C ++, C #, Java, Objective C, Object Pascal. There are also scripting languages like Perl, Python, PHP and Ruby.

It may be preferred that the software module is only installed once on the server hardware and processes all client requests, whereby it is executed only in a single instance, in particular if all clients accessing the server have the same version of one Use server software. Thus, existing resources can be optimally utilized. For the purposes of the invention, the term "client" designates in particular a group of users who are authenticated for access to the server, in particular the server programs. It may be preferred that at least two clients access the server. The requests are preferably made using the hypertext transfer protocol. That is, the clients, in particular the users make the requests in HTTP, preferably via a web browser, which is installed on a user's computer.

In order to ensure the mutual impairment of the clients (ie the user) as well as the access to only their own data, it must be determined with each request to the server software (a) to which client the request is to be assigned, (b) by reconfiguring and rebuilding the channels to the external systems holding status and persistence data, pruning the data area of the server software with the data of the client, and (c) retrieving the client's data from the data area of the server software after execution of the request be removed.

It may be preferred that the once established channels be maintained to the external systems for a certain period of time, even if they are not currently needed during the period.

Request-based server programs are preferred in which status and persistence data are kept in external systems such as database management systems or key-value storage.

In addition, servers are preferred in which the client does not have their own access to the operating system level and can not install or modify software themselves since only one instance of the software is preinstalled for all clients.

In particular, the client does not have the ability to access data areas of other tenants because it does not have access to the operating system or file system level. This type of deployment is suitable for clients with uniform requirements. For the purposes of the invention, the term "requirements" refers in particular to processes or tasks that are to be carried out with the aid of the server programs.

It was completely surprising that the preferred method considerably simplifies the maintenance of the server, in particular the server programs, and reduces the maintenance effort. Essentially, only the installation of the software module and a single installation of the request-based server program need to be maintained. Resource consumption (memory, persistent storage, computational power) is very low and less than the methods or devices disclosed in the prior art Surprisingly, no own resources are blocked per client. The advantage lies in the fact that the resource consumption of, for example, 100 users of 100 different clients (ie per client 1 User) is equal to that of 100 users of only 1 client.

It was further surprising that the number of clients that can be accommodated per physical server computer can be increased and above that of the prior art.

It is preferred that for each request to the server software, the data areas of the software are populated with the data of the affected client.

Request-based server programs preferably use external systems for holding status and persistence data, which preferably include relational database management systems (RDBMS) and key-value stores. A database management system (DBMS) is preferably the software that is installed and configured for a database system. The DBMS defines the database model, has to secure a large part of the requirements listed below and decisively decides on functionality and speed of the system. A relational database is used in particular for electronic data management in computer systems. In the context of the invention, the associated database management system is referred to in particular as a relational database management system or RDBMS (Relational Database Management System). For querying and manipulating the data, preferably the database language SQL (Structured Query Language) is used. The relation represents a mathematical description of a table. Operations on these relations are determined by relational algebra. Relational algebra is thus the theoretical basis of SQL.

A relational database is preferably a collection of tables (the relations) in which records are stored. Each row (tuple) in a table is a record. Each tuple consists of a set of attribute values (attributes = properties), the columns of the table. The relation schema specifies the number and type of attributes for a relation.

To address and access these systems, server programs have configuration parameters that indicate which communication channels - such as sockets on the local hardware or TCP / IP and IP addresses in the network - and what access data should be used.

It is preferred that all status information and all persistence data of the clients or clients of a server be kept in these external processes, so that a request-based server program as such may be considered non-stateful with respect to a single request if the configuration parameters are for access to the stateful systems with each request are variable.

The software module is preferably tuned to the server software and preferably installed at the time of installation of the server software.

The invention further relates to a computer program product on a computer-readable medium for supporting request-based server programs in a client-server software system having a plurality of program elements and means for carrying out a preferred method.

Moreover, the invention relates to a computer system for supporting request-based server programs in a client-server software system comprising means for performing a preferred method and means for implementing a client-server software and hardware system. The above representations of the method are analogous to the computer program product and the computer system to apply.

In the following, the invention will be explained with reference to an example, but without being limited thereto.

Example 1: Procedure of a request

The software module is called on each request before the execution of the actual server program to (a) extract from the request the information about which client the request is and (b) adjust the configuration parameters of the server software and to reload the current request for the data areas provided to the tenant within the external processes for status and persistence data. For this purpose, the software module has an assignment table of client identifications to configuration parameters. In the final step, the software module then passes (c) control to the server program, which starts execution as if the connection to the external systems had not occurred. An essential feature of the invention is the fact that the server program modified by the software module does not have to "know" the concept of multi-client capability. In the context of the server program, there is no knowledge that it is used for several clients in turn. This makes it possible to adapt already available standard software so that it can be used for several clients.

ID

To identify the client, different host names (eg subdomains) can be used in the example of HTTP requests. If the hostname
mandant1.provider.de and
mandant2. bereitsteller.de
For example, if both are associated with the server machine that installed the server program and the software module, the software module may use the HTTP request addresses
http://mandant1.bereitsteller.de/login.html and
http://mandant2.bereitsteller.de/login.html
derive the affiliation to the respective client and set the configuration parameters accordingly, for example to display a login page as in the above case.

The authentication of the corresponding user in the example above and in general is not a task that the software module must perform. After the client has been identified, this function can be carried out quite regularly by the server program itself, since at this point in time it already has access to status and persistence data which, as a rule, corresponding authentication features (user names, passwords, etc.) for the user context within the user context Server programs include.

assignment

To get from a tenant identification to the set of configuration parameters, the software module has an allocation table. The assignment table contains per IP address or socket paths as well as access data.

• mandant1:
  • mysql:
    • database: „mandant1“
    • host: „192.168.0.1“
    • username: „mandant1“
    • password: „abc123“
  • memcache:
    • host: „192.168.0.2“
    • key_prefix: „mandant1_“
• mandant2:
  • mysql:
    • database: „mandant2“
    • host: „192.168.0.1“
    • username: „mandant2“
    • password: „xyz123“
  • memcache:
    • host: „192.168.0.2“
    • key_prefix: „mandant2_“

As an example, the software module can use the following assignment table:

• mandant1:
  • mysql:
    • database: "client1"
    • host: "192.168.0.1"
    • username: "client1"
    • password: "abc123"
  • memcache:
    • host: "192.168.0.2"
    • key_prefix: "mandant1_"
• mandant2:
  • mysql:
    • database: "client2"
    • host: "192.168.0.1"
    • username: "client2"
    • password: "xyz123"
  • memcache:
    • host: "192.168.0.2"
    • key_prefix: "mandant2_"

If the client is determined by extracting the tenant identification, the software module can now establish the connection to the external processes MySQL (an RDBMS in this example) and Memcache (a key-value store in this example) and status and persistence data are available to the server program accordingly.

Transfer of control

After successfully establishing the connection to the external processes, the software module transfers control to the actual server program. Depending on the technical characteristics of the server program, this happens with mechanisms provided by the server program or programming libraries and frameworks used. The software module usually takes the place of a decorator here.

It is advantageous if the server program is present in the source text or has interfaces by means of which the addressing of status and persistence systems can be controlled. Since request-based server programs today mostly use scripting languages that are not compiled, this is generally true.

It is further preferred if the server program keeps all status and all persistence data in external systems (RDBMS, key value stores, etc.) (ie stores). Since request-based server programs are nowadays mostly designed to be used in highly available and load-distributing environments, this is generally also fulfilled.

If the server program defines process-local volatile data areas (caches, etc.), the software module can empty them accordingly before each request. In general, for high availability and load balancing, server programs can also be configured to use external systems for volatile data.

Differentiation to the state of the art

Application level virtualization uses a hypervisor, just like regular virtualization. However, unlike regular virtualization, there is no full operating system environment available for the server program, only the parts of the environment that the program must necessarily access in order to function.

Application-level virtualization is more cost-effective in terms of resource consumption, but at best only as low as the addressed parallel hosting, whereas the mechanism in the disclosed invention has only the resource consumption of a single instance of the program.

• 1 Stand der Technik - Interaktion zwischen Clients und einem anfragebasierten Server-Programm
• 2 Stand der Technik - Interaktion zwischen Clients und einem anfragebasierten Server-Programm innerhalb einer virtuellen Gast-Umgebung
• 3 Stand der Technik - Interaktion zwischen Clients und einem anfragebasierten Server-Programm innerhalb einer gemeinsamen Betriebssystem-Umgebung
• 4 Anfrage an ein Server-Programm mit einem bevorzugten Software-Modul
• 5 Darstellung eines bevorzugten anfragebasierten Server-Programms
• 6 Bevorzugte Herstellung eines zu einem Mandanten passenden Kanals zu einem externen System

The state of the art and the invention will be explained below by means of examples, but without being limited thereto. Show it:

• 1 Background Art Interaction between clients and a request-based server program
• 2 Background Art Interaction between clients and a request-based server program within a guest virtual environment
• 3 Background Art Interaction between clients and a request-based server program within a common operating system environment
• 4 Request to a server program with a preferred software module
• 5 Representation of a preferred request-based server program
• 6 Preferred production of a client-matching channel to an external system

1 shows the state of the art, in particular an interaction between clients and a request-based server program. They are different users 1 shown with her web browser 2 over the internet and http 3 with the request-based server program 5 to interact. The server program is installed on a server computer with operating system environment 4 where it occupies hardware resources such as working memory, persistence memory and computing power 6 , The operating system services and programs required to operate the server machine 7 also occupy resources 8th ,

2 exemplifies the state of the art, in particular an interaction between clients and a request-based server program within a virtual guest environment. The users of different clients 1 / 1 . 2 , each with their web browser 2 over the internet and http 3 with the server program installed for the respective client within the virtual guest environment provided for the respective client 9 which in turn interact within the host environment 4 is performed. Each virtual guest environment 9 occupies own resources 6 in the host environment 9 , The operation of the server machine, the host environment 9 provides necessary operating system services and programs 7 also occupy resources 6 ,

3 shows a prior art, in particular interaction between clients and a request-based server program within a common operating system environment. There are different users of different clients 1 / 1 . 2 shown, each with their web browser 2 over the internet and http 3 with the server program installed for the respective client 5 interact within the common operating system environment on the server 4 is performed. Each installed server program uses its own resources 8th , The operating system services and programs required to operate the server machine 7 also occupy resources 8th ,

4 shows a request to a server program with a preferred software module. They are the users of different clients 1 / 1 . 2 shown, each with their web browser 2 over the internet and http 3 with the single server program installed for all clients 5 interact within the common server operating system environment 4 is performed. The server program 5 occupies own resources 6 but only simple - not separately for each client as in the previous examples. The to the operation of the server computer 4 necessary operating system services and programs 7 also occupy resources 8th ,

5 represents a preferred request-based server program. It becomes the various instances of a request-based server program 5 and abstracting from the property, whether they are installed in parallel or virtualized. These access via permanently configured communication channels 10 on an external system 11 , the status and persistence data 12 holds, too. It is also shown that the persistence data 12 as is customary today, divided into different areas for different clients.

6 illustrates a preferred creation of a tenant-matched channel to an external system. It is shown as a preferred embodiment of the module 13 is used to one and the same installation of the server program 5 Depending on the request, the appropriate channel for the respective client 10 to the external system 11 to provide. The external system 11 also provides status and persistence data in this case 12 available only for the client's intended area, analogous to 5 ,

LIST OF REFERENCE NUMBERS

1 / 1.2

user

2

Web browser

3

Internet and / or HTTP

4

Server computer

5

request-based server program

6

resources

7

Operating system services and programs

8th

Resources of operating system services and programs

9

virtual guest environment

10

communication channel

11

external systems

12

Status and / or persistence data

13

module

Claims (13)

Method for supporting request-based server programs in a client-server Software system, wherein a module is present on a server, in particular a server software, and the module is called upon request of a user to a request-based server program, and in turn extracts from the request the identification of the client environment to which the request is directed, wherein the module, by means of a mapping table, assigns a set of configuration parameters to the query based on the tenant identification, and adjusts and reloads the configuration of the request-based server program so that the request for the data areas provided to the user within the external processes for status and persistence data are accessed, whereby the software module after the connection passes the control to the server program, which then the processing of the request begins as if the reconfiguration never took place. Method according to Claim 1 , where the server program is installed on a server hardware and is constantly running there. Method according to Claim 1 or 2 in which the requests are processed by means of server hardware, in particular main memory, persistent memory and computing power. Method according to one or more of the preceding claims, wherein the software module is installed only once on the server hardware and processes all client requests. Method according to one or more of the preceding claims, wherein the request-based server program is connected to systems for holding status and persistence data. Method according to one or more of the preceding claims, wherein the systems for holding status and persistence data comprise relational database management systems and / or key-value memories. Method according to one or more of the preceding claims, wherein the allocation table IP addresses, socket paths and / or access data, in particular user name and password includes. Method according to one or more of the preceding claims, wherein requests are made using the hypertext transfer protocol. Method according to one or more of the preceding claims, wherein a web browser is installed on a client computer. Method according to one or more of the preceding claims, wherein for each request to the server software, the data areas of the server software are populated with the data of the affected client. Method according to one or more of the preceding claims, wherein already available standard software can be adapted so that it can be used for several clients. A computer program product on a computer-readable medium for supporting request-based server programs in a client-server software system having a plurality of program elements and means for performing a method according to the Claims 1 to 11 , A computer system for supporting request-based server programs in a client-server software system comprising means for performing a method according to the Claims 1 to 11 , and means for implementing a client-server software and hardware system.

Images