Sometimes I visit big vendors’ presentations where they talk about their new servers. The audience — systems administrators, engineers, technicians — are listening very carefully, trying to grasp every word. They like to hear that the new server, equipped with the new CPUs, can achieve 20% or 40% bigger performance, “at the same price tag”.
Come on, guys. Don’t strain your ears. Isn’t it time for something new?
So, Intel recently announced their plans to acquire QLogic, the manufacturer of InfiniBand hardware and software.
I just wanted to compare the two major InfiniBand manufacturers, QLogic and Mellanox, by their sizes. Here is the graph:
If you are a pro in corporate finance, you can draw your own conclusions from these bars. In general, QLogic has a twice bigger head count, and they have sold about twice more products and services during their last financial year.
But what if you wanted to compare these two companies to Intel?
The February 2012 issue of the “Communications of the ACM” features an article by Bryce Allen et al., “Software as a service for data scientists” (or see PDF).
The main idea of the article is as follows. The amount of research data is growing exponentially. (Can we say: “the quantity of data is turning into a new quality”?) Then, one day you need to transfer a big amount of data from one data storage facility to another. The data is in files that can be either big or small, and there are many files — more than you can handle manually. Moreover, some files don’t need to be transferred because they are already present on the target system. Also, you are not sitting in either of the facilities: you are at home, in front of your notebook. And you want to start the transfer and go to bed, so that every failed transfer is retried automatically.
So, the solution that the Globus Online team came up with is a sophisticated system that launches the transfer directly between the two facilities, using the GridFTP protocol. Your notebook is only a console to control the process with a Web interface (or a command line). Rather elegant, indeed.
Back to the good old days when I worked as a system administrator, I remember that I wanted to implement a Lustre file system for the two small clusters that we had (64 and 20 nodes, respectively).
However, after learning that it requires Linux kernel patches, seeing how many people ask for help on the mailing list, and finding a 500-page manuscript on installation and basic management, I decided to wait. Seems that Lustre was mostly for big corporations at that time. Besides, it was not compatible with Microsoft Windows-based clients which was a crucial requirement for us.
Nevertheless, the installation and management difficulties are likely to diminish with the advent of Chroma, the software by Whamcloud.
Russian supercomputing company, T-Platforms, was quoted on 23 December 2011 by “Computerworld” as claiming to build a 10-petaflop system.
Okay, it was me who wrote “threatens”. T-Platforms didn’t state it so. What is more interesting is that their website still (as of 13 January 2012) doesn’t contain any reference to that claim. (UPD from 14 February 2012: here is the link, dated by 21 December 2011. Probably I missed it?)
However, the institution where the machine would be installed — the Moscow State University — did have this information published one day earlier on its affiliated website, “Supercomputers.ru”
