New Impulse-50 (DPC Projects of creation and upgrade)

2017: A data processing center for ensuring work of retail chain stores Duck-bill

Tasks

With growth of business and increase in quantity of sales points before company management "New impulse" there was a task: ensure continuous, smooth functioning of retail chain stores Duck-bill and safety of information in case of hardware failures, program failures and human errors.

The system of preorders imposes high requirements to the information system (IS) of the company as it is necessary to provide support of a large number of transactions between shops and a warehouse storage: transfer of the order from shop for a warehouse in real time, control of existence of goods in a warehouse, control of passing of the customer order, etc. As all customer information, goods and orders is consolidated in the centralized database, failure in operation of the IC can lead to a complete stop of business, i.e. to heavy losses. As processing system and data storage about clients and orders of New Impulse company the SAP R3 system was used.

Growth of number of shops and increase in a flow of clients in translation into the IT language means increase in volumes of the processed data. It was predicted that the number of orders will average not less than 10 thousand a day. Besides, it was going to implement resource-intensive software for the analysis, forecasting and work planning of big retail chain stores, and the existing computer system had no reserves for performance improvement. The solution of these tasks required upgrade of the computer system of the company.

To provide support business processes, the company management "New impulse" made the decision about construction a complex of buildings in the Area of Yuzhnoye Butovo. Here the warehouse storage, office spaces and data processing center (DPC), supporting work of retail chain stores in To Moscow, Zelenograd and Solnechnogorsk had to be placed.

Choice of the partner

For the choice of the general contractor of the New Impulse project held closed competition in which won a victory there was a Jet Infosystems company which, according to customer representatives, "had necessary experience of creation of the similar systems, highly skilled personnel, competent project management".

Thus the Jet Infosystems company became the partner of "A new impulse" in creation of DPC, and the tasks to design and construct engineering infrastructure and the computer system for work of ERP- and other enterprise information systems were set for her specialists.

Work progress

Specialists of Jet Infosystems undertook construction management of DPC, exercising control of all stages of the installation and construction works performed by contract organizations and preparation of a complex for commissioning. The DPC was going to be placed in two buildings. At the time of the beginning of cooperation one of them was already constructed, and for the second the platform under the base was prepared.

Requirements to construction contract organizations on creation, placement and finishing of server rooms in buildings taking into account features of a configuration of the premises intended for DPC, the requirement to the sizes and geometry of entrance doors, design features of a floor, materials used for finishing were formulated.

After completion of construction works, specialists of Jet Infosystems company started installation of engineering infrastructure of DPC and design of the computer system.

Engineering infrastructure of DPC

During creation of engineering infrastructure of DPC works on design, installation and start in operation of the following subsystems were performed:

• conditioning,
• gas fire extinguishing,
• mortgages and cable channels,
• regular power supply, including a distribution network of power supply of server rooms,
• the structured cable subsystem,
• and also number of additional subsystems, in particular, general power supply and emergency lighting.

Technical details:

• Subsystem of mortgages and cable channels. Equipment of Niedax and Vergokan. A system total length — 216 m.
• Cable subsystem. Equipment of The Siemon Company. System capacity: 816 copper ports; 544 optical ports.
• Subsystem of regular power supply. Equipment of Merlin Gerin, Chloride Group PLC. Two UPS complexes constructed with reservation according to scheme N+1, total power — 380 kW.
• Conditioning subsystem. Emicon equipment. Four complexes constructed with reservation according to scheme N+1, total power on cold — 370 kW.
• Raised floor. Uniflair equipment. Total area — 240 sq.m.

Computer system of the SAP R3 system

As information, critical for business of the company, is stored in the SAP R3 database, specialists of "A new impulse" formulated requirements to the computer system as follows: it should provide, first, continuous and smooth operation of retail chain stores, secondly, safety of customer information and orders in case of refusal of the equipment and also in a situation of software or human failures. The configuration of the equipment had to have reserves for performance improvement and reliability augmentation in the future.

By drawing up the commercial offer on creation of the computer system specialists of Jet Infosystems analyzed:

• parameters of functioning of the existing data processing complex;
• marketing plans of New Impulse company for previous years;
• business development influence (increase in number of outlets of network, possibility of expansion of the range of products, etc.) on information infrastructure of the company;
• available capacities of DPC, taking into account perspectives of development of the company.

Besides, possible options of placement of the equipment in new buildings and perspectives of upgrade of DPC in the future were considered.

The offer to design a geographically distributed failover cluster of SAP R3 became result of work, such solution had to most full to meet requirements of New Impulse company regarding performance, availability and scaling of a system.

The SAP R3 system includes the server DBMS Oracle and application servers. The DBMS server works with the data placed on two external disk arrays (see fig. 1).

For ensuring continuity of work of an information system in case of refusal of the equipment it was offered to duplicate key components of IT infrastructure.

Fig. 1 Scheme of a cluster SAP R3 of New Impulse company

For effective access of servers to disk resources the fiber optic storage area network – SAN — by the principle of dual-fabric on four Fibre Channel Brocade SilkWorm 3800 switches was designed and built.

The components of a complex located on different platforms are connected among themselves by optical links:

• two communication lines are intended for a local network and network of intercluster interaction;
• 4 communication lines — for SAN.

For reliability of construction all communication lines were separated to two physically different canals laid in the independent ways.

In the mode of normal work of a complex all clients are connected to application servers. The special process manager of SAP R3 which is performed on one of servers evenly distributes loading, connecting new users to the server which is least loaded at present. The Oracle DBMS server provides the data necessary for work on requests of application servers, and receives results of execution of operations. Communication between servers is provided by means of Ethernet technology. All these systems are stored in files of the database on HDS 9980V disk arrays. On arrays data and other servers are placed.

All equipment which is a part of the computer system has redundancy and the built-in diagnostic aids. Therefore failure of separate elements will lead only to possible degradation of technical characteristics of a complex in the set limits, but not to idle time in general.

In case of refusal any of servers, the disk array, tape library or the equipment of SAN the following options of work are possible:

• Use of a failsafe cluster configuration allows, in case of refusal a primary server of DBMS, automatically, without intervention of the administrator to perform transition to a standby server.
• At a total failure of one of application servers all loading will be assumed remained, at the same time only the current sessions of clients will be lost.
• In case of breakdown of the main HDS 9980V disk array reserve is provided. On it using the software of Veritas Volume Manager the mirror copy of data from the first array is created that allows to increase, on the one hand, the speed of data reading and performance of operation of applications, and on the other hand, to avoid a system stop at failure of one of arrays. At unavailability of one of mirror copies of data work will go with the second without any breaks or stops. After maintenance of the refused array data on it will be synchronized with the work copy, and a system will continue work in a complete configuration.
• The existing SAN configuration uses duplication of an access path from servers to storage systems. Thanks to Veritas Volume Manager in case of refusal of one of ways (failure of the FC switch, break of the optical cable, failure of the FC interface on the server or the disk array) all data stream will be directed on a reserve way.

Specialists of Jet Infosystems company organized carrying out backup procedure online, using technology of creation of "instant pictures" (snapshots). This scheme allows to do backup without stopping of operation of applications, it reduces time of carrying out copying and does not lead to reduction of system performance.

For creation of "instant pictures" the software of Veritas Volume Manager is used.

Service maintenance

The equipment of the computer system is on service support of Jet Infosystems company. The hot line works, visits of the specialist of the Service center for performing diagnostics and rescue and recovery operations are performed, planned modification (reconfiguration) of the serviced equipment and software is carried out.

Basic elements of a complex are on the round-the-clock support according to the Operational Support 24х7 program which is intended for warning of emergence of non-staff situations, minimization of cases of abnormal failures and fast recovery of functionality of the high-speed equipment and software which is in commercial operation. The program provides the guaranteed term of complete recovery of functionality of a system and also support of all system entirely, but not separate elements of the equipment and software.

Control of shops of Utkonos network is exercised from a uniform data processing center. The complex of the proposed technical solutions and organizational measures allowed to ensure safety of information and continuous and smooth operation of retail chain stores. As a part of a complex:

• geographically distributed cluster,
• duplication of key components of IT infrastructure,
• creation of SAN,
• organization of carrying out backup "on-line",
• support of the computer system according to the Operational Support 24х7 program.