The Lyon metro; The Lille, Le Mans, or Melbourne tram systems? Keolis helped to design these transport systems and operates them on a daily basis. As the largest urban transport operator in France and the leading tram operator worldwide, the company operates 85 urban transport networks and covers 74 French counties with interurban links. In 2012, Keolis transported 2.2 billion passengers and achieved a turnover of 4.9 billion euros. Its philosophy? Placing innovation at the heart of its approach to modes of transport, associated services, and information technology. With this philosophy in mind, Keolis Orléans wanted to strengthen the efficiency of its IT systems, particularly with regard to operations and production. One of the organization’s objectives was to make increasing use of VDI technology through VMware View 5.1, which is already partially deployed in its infrastructure. Keolis Orléans wants a solution that is able to anticipate business evolution: the number of virtualized desktops will not be known in advance, and the volume of associated data will be even harder to predict. A high-performance, scale-out array is therefore needed, one that can be easily integrated into the highly virtualized IT environment. After keeping a close eye on technological developments, Keolis Orléans ultimately selected the EMC XtremIO array, as it meets the specifications perfectly. The arrival of this new class of All-Flash storage systems on the market redefines what can be offered in terms of performance. “When we chose an XtremIO array 18 months ago, it was the most advanced product on the market, with the most functionality at a reasonable cost. We’re actually talking about a device built to perform Flash by design, and not a variety of disk arrays to which you would just add SSDs,” explains Olivier Parcollet, Information Systems Architect at Keolis. The array combines the use of Flash technology as a storage medium with systematic use of data deduplication to minimize Flash wear and maximize capacity utilization. Because the architecture is based on a scale-out design, performance increases with the number of nodes in the storage cluster. The aim is also to ensure that the array automatically optimizes itself in such a way that no fine-tuning or manual resource allocation is required.
⇩ Register ⇩