GigaSpaces' platform provides a basis for scaling appliations written a spaces based architectute, using the notion of a processing unit (quote from previous coverage):
In Space-Based Architecture, a Processing Unit represents the unit of scale and fail-over of an application. It normally contains all of the application services and middleware components that have a tight latency/runtime dependency. It encapsulates those services under a single container (Processing Unit) and maintains consistent scaling and fail-over semantics in a generic fashion to all of those components. A failure event, for example, will automatically trigger a recovery process of both the middleware components (messaging, data grid) and the business logic associated with it. In this way we avoid partial failure or inconsistent behavior resulting from the fact that a failure event happened, messaging system started to deliver events but the application service is not yet ready to process them. From a latency perspective, the encapsulation of all those components in the same run-time container reduces network overhead, because they interact purely in memory. Scalability becomes as simple as adding more processing units. In other words, there is no need to separately scale the data, business logic and/or messaging tiers.InfoQ asked Geva Perry about the business strategy behind offering their flagship product for free:
It’s good for start-ups, good for our enterprise customers and good for GigaSpaces’ business. For start-ups it’s all about the ability to easily scale-out their application as their business grows, with little or no code changes. Nati Shalom wrote a great blog post with his impressions from the last QCon in San Francisco that touches on why start-ups don’t deal with scalability from day one. Historically, scaling is complex, so companies at an early stage prefer to postpone the issue because they are more concerned with time-to-market. But with GigaSpaces, that complexity is eliminated – so they don’t have to make that trade-off. The only issue that remains then is cost. With the Start-Up Program we’re eliminating that issue as well. For our non-start-up customers it is good because higher adoption of GigaSpaces creates a bigger community of GigaSpaces users, eventually leading to a more robust and richer product. For GigaSpaces, it makes a ton of business sense. We have built a very successful enterprise business, and our customers include the world’s largest banks, telco companies, government agencies, web companies and more. When we look at our revenues, a tiny fraction comes from companies with less than $5 million in sales. With this offer, we can gain those customers early on, and when their business grows, there’s an opportunity for us to make money. We’re sharing the risk with them basically saying “if you don’t make money, we don’t make money”.On how applicants will be auditted to ensure they really are < 5M in revenues, Geva replied:
There is a self-policing dynamic here, which is working quite well, so we really don’t need to audit. First, companies with significant sales or VC funding are not likely to knowingly violate a contract, so we don’t think there is a big risk of that happening. Even if it somehow happens unknowingly, it’s not really a problem, because at some point -- if anything of consequence is happening with the application built on GigaSpaces – the customer will want 24/7 support with demanding SLAs. In order to receive that, they will need to come to us. In addition, we have real people looking at the applications, so if there is any application that we think is not really a fit for the start-up program (such as academic research), we simply send an email and ask what’s up.The clustering & caching space has a number competitors including Coherence (not owned by Oracle), Gemstone, and Terracotta which open sourced its Java clustering solution one year ago. Geva was asked if going free for a certain market segment was the first step towards eventually open sourcing their product, to which he replied:
I think that what we’re seeing increasingly is the notion of hybrid models. There is a spectrum of models from one extreme of free open source to the other extreme of everyone-pays closed source. Then you have things in the middle such as commercial open source (e.g., JBoss, MySQL). Open source has the benefit of fast adoption because it lowers entry barriers (for users) and fosters a large community, which is goodness. But building a successful business around it is tough, and so far we’ve only seen clear-cut examples of commercial success in areas where the product category is mature, even commoditized. GigaSpaces is dealing with innovative technology and has a great enterprise business, so we needed a different model. We follow very closely what companies such as Atlassian and Jive Software are doing. We’re striving for what we call internally a “progressive model” (paraphrased from “progressive tax”). Meaning – customers who use our product to make money should pay for it. Period. And that’s where our revenues come from. But there are other audiences: academic, start-ups, evaluators, who should have free access to the software. In addition, this is part of a larger effort. We open sourced our OpenSpaces framework, which is our flagship API based on the Spring Framework. Very soon we’re going to launch OpenSpace.Org – a true community site that allows code contributions. We’re giving full access to our documentation online and are working with open source projects through our free Community Edition. These are just a few examples of how we can be an “open company” without giving away the family jewels for free. So to answer your question, we are not going to be free open source in the foreseeable future, but we are working hard on being open and community-friendly. As Matt Asay describes beautifully in his blog, we’re following a philosophy of abundance, not scarcity.Vertical Acuity, an Atlanta, Ga.-based startup has joined the GigaSpaces Start-Up Program. Brent Walker, a co-founder of the company who leads engineering said: “The GigaSpaces Start-Up Program comes at a perfect time for us. We wanted to benefit from the scalability and performance that the GigaSpaces product provides, but having raised only seed money, we can certainly use the economic benefits of this program.” Vertical Acuity is developing a Web service that will provide their customers with market and competitive information based on online activity in a particular industry, such as automotive or insurance.