Every once and a while, someone on the Beowulf Mailing List asks, Where can I find the Beowulf software? Or, What makes something a Beowulf? Each time after some discussion or wrong notions and answers I pull out my stock answer and and webpage link. As a service to the community and in an effort to keep the truth alive, I will explain, yet again where the name came from and why there really is no such thing a Beowulf Cluster. But first, an interesting historical anecdote.
At some trade show in a galaxy far, far away, I was standing near a large vendor’s booth where a sales person was explaining the name Beowulf to some would be customers. “Beowulf was a copy of the Microsoft Wolfpack software. They chose that name so it would seem like Wolfpack some how”. I have always found this comment curious since the Beowulf Project was started in 1993 and the first time Microsoft Cluster Server (aka Wolfpack) appeared was 1996. Well, OK so the sales person was off by a three years. I’ll also skip the fact that the Beowulf project was designed for HPC and MS Wolfpack is for load balancing and fail-over clustering. I’m not one to split hairs. I also note that HPC Windows Compute Cluster Server 2003 was released in 2006, so baring any relativity effects, I’m guessing someone just needs a new calendar.
In any case, such folklore is why I continually try and set the record straight. Indeed, the whole story has been right here in Linux Magazine since 2003. That is 2003 as we all know it. In any case, I figured a story written by Tom Sterling, the guy who started the project, wrote the books and papers about it, might have more truth to it than some sales drone.
As a special bonus, I’m going to supply the key paragraphs below. By the way, Jim Fisher was the NASA project manager who funded the Earth and Space Sciences (ESS) project at the Goddard Space Flight Center (GSFC).
Then one afternoon, Lisa, Jim Fischer’s accounts manager, called me and said, “I’ve got to file paperwork in 15 minutes and I need the name of your project fast!” or some words to that effect. I was desperate. I looked around my office for inspiration, which had eluded me the entire previous month, and my eyes happened on my old, hardbound copy of Beowulf, which was lying on top of a pile of boxes in the corner. Honestly, I haven’t a clue why it was there. As I said, I was desperate. With the phone still in my hand and Lisa waiting not all that patiently on the other end, I said, “What the hell, call it ‘Beowulf.’ No one will ever hear of it anyway.” End of story.
But the other truth is I didn’t actually name Beowulf, the computer. I only named Beowulf, the project. Someone out there in the land of the press coined the term “Beowulf-class system,” not me. I would love to know who it was. That’s the real irony: I get the credit for naming Beowulf clusters and actually I didn’t do it.
Initially there no such thing a Beowulf Cluster. The name seemed to stick for clusters that were built similar to the first Beowulf at GSFC. Later on, in the book “How to Build a Beowulf” by Thomas Sterling, John Salmon, Donald J. Becker and Daniel F. Savarese, the following definition was given for a Beowulf cluster:
.. a collection of personal computers interconnected by widely available networking technology running anyone of several open source Unix-like operating systems.
The move from a project to a cluster definition just kind of happened, but still the definition is rather loose. I like to look at it as more of a concept or methodology. The idea caught on, however, and the rest became history.
As a keeper of the flame as it were, I am also going to provide brief answers to some common misunderstood questions about “Beowulf.”
- What is a Beowulf Cluster? See above.
- Can I build a Beowulf Cluster with Microsoft software? Not according to the definition above.
- Where do I get the Beowulf software ? There is no secret software. There is Linux, MPI, networking, job schedulers and other tools. You can get cluster distributions that include all the software, but it is not called Beowulf software.
- What is Beowulf? A story from English literature. It was made into a movie that probably used Beowulf style rendering clusters to do the animations.
- And, as an added bonus, If I build a Beowulf cluster will all my programs run faster? No. There is no free lunch.
And thus is how it came to be. Next time someone tries to foist some vague notion of how Beowulf was named you have the answer. You too can become the defender of the faith.
Douglas Eadline is the Senior HPC Editor for Linux Magazine.