Consolidation of several datacenters into a common hyperconverged solution for PNS

První novinová společnost a.s. is the largest distributor of press on the market. From more than 450 suppliers it supplies print, electronic media and additional goods to nearly 17,000 sales outlets across the country. It supplies the densest network of sales outlets in Europe. It distributes 9 national and 70 regional newspaper titles, more than 450 publishers and 33,000 tonnes of newspapers a year to the free market.

The gradual development led to the fact that the IT infrastructure of the First Newspaper Company was spread over a total of three data centres, either owned or leased from traditional providers. Managing such an environment was not easy and its demand on human resources, IT administrators was considerable.

With the merging of several IT environments under one "roof" of PNS, the time was ripe to consolidate a large number of hw different systems from different vendors. To reduce the number of managed devices and simplify their management to free the hands of IT administrators to solve common operational problems of users as quickly as possible.

Main project objectives:

• Unification of different hw maintenance periods
• Reduce the number of data centres needed
• Optimizing the number of licenses
• Preserving the VMware vSphere platform
• Enabling sharing of computing resources
• Possible growth but also further consolidation
• Maintain resilience of critical parts of IT infrastructure against failure - do not reduce availability of e.g. data storage layer
• Simplify the administration of IT infrastructure operations
• Enabling easy migration of virtual servers from all existing data centres

We first deployed an automated tool for collecting and analyzing performance parameters of servers, storage and network traffic (LiveOptics from DELL EMC) in all PNS data centers.

After scrubbing some of the cloud data center data and thoroughly analyzing all the findings, we arrived at the final configuration. Just four physical servers and two Ethernet switches were enough to replace all the hw needed to run the three existing data centers.

The servers will have sufficient processing power and memory to run all the necessary virtual servers and we used the internal disks in these servers as data storage. To make the disk capacity highly available, we selected one of the world's most widely used hyperconverged systems running on the VMware vSphere vSAN platform - DELL VxRail.

Hyperconvergence here has achieved all the goals we set at the outset:

• Warranty periods for all IT infrastructure hw were unified. We reduced the number of data centers from three to one (not counting the DR site). We kept the VMware vSphere server virtualization system, long used by administrators, and thanks to it and VSAN, we not only shared server power for all virtual servers in the cluster, but also provided them with highly available, scalable data storage.
• The number of nodes in VxRail can be continuously increased and, if necessary, decreased. The entire IT infrastructure is highly resilient to simple hw failures not only in the servers but also in the data storage, similar to a classic IT infrastructure with a pair of midrange data stores with transparent failover.
• The administration of the IT environment has been simplified simply by the fact that only one of the three data centres remains. In addition, the administration of the hyperconverged VxRail system is carried out from a single integrated tool.