High Performance Computing Cluster aka Supercomputers, after working for 4 years in this domain( Administration of HPC Cluster) I could say, this is not as challenging job as its sounds (Supercomputer).Its all about configuration, some people take days, some takes months and some could take years depending on knowledge and experience but ultimately every one can setup these clusters.
HPC Clusters are basically individual machines connected with network and have softwares installed in client server way. There are basically two ways of setting up cluster, one is to use management softwares in which some are proprietary (like InClus, HP CMU) and some are open source (like Rocks, XCat). Second way is to setup cluster manually. In this tutorial we will use the manual approach, reason is it will give us better insights of HPC configuration.
To understands these tutorials you should have hands on experience on Linux Administration, Network Protocols and Python Programming Language(Could Benefit). We will cover following topics in details:
1. DHCP ( Dynamic Host Configuration Protocol )
2. Network Boot ( TFTP )
3. NFS ( Network File Sharing Protocol )
4. Torque Scheduler ( Xeon Phi and GPGPU Configuration )
5. Ganglia ( Cluster Monitoring Tool )
6. OFED ( Infiniband Software )
These 6 things is all you need to setup manual cluster, we will go through all these topics in details in coming tutorials.
To all the security professional who are following this blog, this topic is really important if you are working on some password cracking algorithm, brute force, DOS or anything which requires processing power of CPU, by setting up such clusters at your home you could execute program in parallel( using processing power of more then one machine at the same time).
Follow us on Facebook.
Thanks
Abhishek.
Networls 