We know your needs

The success of teams Sie Ladon, is based on having been applied for many years the experience we have transmitted our clients on new developments that we have created.

These developments are not accidental, but have been covered in the time gaps that had, which allowed us to further strengthen the clusters and computing machines we offer.

Extensive testing of our equipment before delivery, temperature control systems for the courtrooms where they are, improving the open source development on Linux, new elements of very high-speed interconnect and low latency, allow us to offer very robust systems .

Today, there are still systems mounted 10 years ago and subsequently extended, which are still used with great advantage. Thanks to the way it solves its European information systems implementations, we can always guarantee future growth without trauma. 

 
The temperature control is achieved by a system co-developed by Dr. Enrique Lomba, Director of the Research Center Rocasolano CSIC and Systems engineers Europeans. With this warning system provides for two levels:

Level A: Since the cluster should not exceed 25 º C when it reaches this temperature, which is considered abnormal in a refrigerated room, the system alerts by email or SMS to the operator of this temperature rise, although still operating normally.

Level B: If you reach 35 º C, the server nodes throws off a script. After a few minutes, shuts himself, leaving a log of the peak temperature. This will safeguard the integrity of the system until it is uprooted by an operator, does not cause data loss in the calculations was performed.

The system is completely autonomous with direct access through IP. It can manage both via the web as an integrated SNMP console alerts. This system was devised mainly because our engineers discovered that in most cases, to leave the light, the servers have UPS power and continued to function normally. However, air conditioning systems, electrical energy when made, are not automatically rearmed and end by producing critical situations. Moreover, breakdowns in air conditioning (cooling gas leak, lack of efficacy in the hot months, breakdown maintenance, etc.) are much more common than in high-tech systems. This is because in many cases the clusters are not in CPD but cooled rooms inappropriately.

CLUSTER CONNECTIVITY

SIE clusters using two independent and internal networks, one for data and one for calculations (improving performance in parallel calculations). In the data network using Gigabit to get good performance in the volume of data to move. In the network of calculation, depending on the client, we can use Gigabit or low latency technologies such as Infiniband or Myrinet, with which significantly improves the parallel calculations. In addition to these two networks, the main server is connected to the university or research center through third Gigabit network card and share this connection with the other nodes.

+Details | +Information