Powers Unfiltered

An entrepreneur’s journey into grid computing and partnering with Microsoft, by John Powers

Powers Unfiltered header image 2

Hedge funds and grid computing, part II

June 11th, 2007 · 1 Comment

So how did the Digipede Network become the grid computing software of choice for hedge funds?

The three primary value propositions that we’ve been emphasizing from the beginning resonate particularly well with hedge funds. For those of you who are new to the Digipede message, those propositions are: Performance, Simplicity, and Microsoft focus. And if you don’t believe me, then go read what hedge fund CTO Paul Algreen is saying in GridToday. Go ahead — I’ll wait.
Now — Let’s look at how hedge funds respond to three bullets from an early slide we’ve used since before we released the first version of our software:

Performance: “The Digipede Network delivers dramatically improved application performance at any scale.”

Lots of people think “at any scale” is some sort of bragging about the ability to manage thousands or tens of thousands or gazillions of compute nodes — and of course it is. But just as importantly, “at any scale” means that we are able to scale DOWN as efficiently as we scale UP. Most hedge funds (even some whoppers) are NOT huge enterprises — they are small-to-medium businesses with some extremely compute-intensive applications. Our hedge fund customers have all found that they can get up and running with a “starter kit” of 10 or 20 nodes, verify the benefits of “dramatically improved application performance” at that level, and then scale up from there — adding more nodes, more applications, and more users as needed.

Simplicity: “The Digipede Network is radically easier to buy, install, learn, and use than competing grid offerings.”

That looks like marketing language, and it is. But before this sentence was ever used in marketing, it was an internal mantra at Digipede that drove literally hundreds of specific design and engineering decisions. (It’s still taped to the wall in our original conference room.) Hedge funds don’t have big IT staffs, and they don’t go for years-long big-budget “grid-enablement” projects. They need something that gets their best ideas for trading and risk management into the market now, not after they figure out how to implement some needlessly complex monstrosity. And they certainly aren’t going to wait while they re-train all their developers to learn new and alien programming paradigms. So hedge funds appreciate how naturally the Digipede Network fits with their current IT environment, and their current development practices.
Microsoft focus: “The Digipede Network is the only commercial grid computing product built entirely on .NET.”

And while THAT one raised some eyebrows when we first came to market (I’ve lost count of all the times I heard that there was “no such thing” as high-performance .NET computing), this has been a great decision for us. Hedge funds work in Windows, period. (OK, of the 10,000 hedge funds out there, I’m probably wrong about one or two. Deal with it.) There’s no room at a hedge fund for a “high-performance computing” group working on an entirely different platform than the “all the rest of the computing” group. Developers at hedge funds use a wide variety of tools and applications — .NET, COM, stand-alone executables, third-party libraries — but everything runs on Windows. Digipede’s early decision to let competitors pursue the false goal of “inter-OS” deployment while we integrated deeply at multiple points in the Microsoft stack has paid off well in this market.

So — it’s no secret. Performance, Simplicity, and Microsoft focus. Simple, but effective — now to spread the word to those next 9900+ funds…

Tags: Grid applications · Growth · Partnering with Microsoft

1 response so far ↓

  • 1 How Digipede became grid software of choice for hedge funds | insideHPC // Jun 11, 2007 at 3:33 pm

    [...] Interesting post from John Powers (who regularly posts insightful comments on an industry I don’t get to see in my day job) on the what it takes to be successful in that rapidly growing segment of the market, enterprise HPC: meet customers where they are. Lots of people think “at any scale” is some sort of bragging about the ability to manage thousands or tens of thousands or gazillions of compute nodes — and of course it is. But just as importantly, “at any scale” means that we are able to scale DOWN as efficiently as we scale UP. Most hedge funds (even some whoppers) are NOT huge enterprises — they are small-to-medium businesses with some extremely compute-intensive applications. Our hedge fund customers have all found that they can get up and running with a “starter kit” of 10 or 20 nodes, verify the benefits of “dramatically improved application performance” at that level, and then scale up from there — adding more nodes, more applications, and more users as needed. [...]

Leave a Comment