Open-source servers are already cheap and secure — a key selling point in the market for Oracle’s (ORCL) NoSQL database — and other third-party providers like Citrix Systems (CTXS) and JBoss offer better performance, higher availability and lower latency.
But these clouds are limited in the other two areas — storage and Web operations. Which leaves the question of where open-source servers need to go next. Microsoft’s SQL Server is the industry standard for database servers, but based on Google’s (GOOG) popular OpenShift container management platform, the market isn’t saturated — and there’s lots of opportunity for open-source developers to provide attractive solutions.
Solutions such as Supernode — a technology developed by Google with the help of Internet service provider Meraki — offer performance-enhancing support at lower prices than Google’s proprietary system. Another contender is OpenSolaris. This serverless software sells for $300, but can scale as low as $15,000, according to Web-hosting services, because the platform is scaled to the compute, storage and networking resources required to support all of the workloads on each instance of the virtual server.
Some may not be interested in these approaches because of the complexity of controlling hundreds of thousands of servers. But Linus Torvalds has said repeatedly that he’s working on a way to build networks of servers — without having to move servers around or reduce high-end computing power. Torvalds calls this “open-source Hadoop distribution for individual servers.”
According to an article in the New York Times, the Eisai Linux systems operating system is today being challenged by a new open-source technology called Newton. Newton is based on vast amounts of unused computer power, along with access to low-cost networking hardware, that these other vendors are not willing to take up.
This means that the goal here is to accelerate the development of organizations that want to start migrating legacy systems out of the confines of proprietary mainframes into the private cloud. Much of that activity is currently going on for open-source developers, although CloudFlare, Box and others don’t offer their own full services or products.
Fortunately, the field is more mature than the issue that was on the minds of many journalists that I spoke with at the Jeff Bezos privacy conference last year. One of my guests was Tom Lawton, chief operating officer of SnapStream, a startup that makes devices for monitoring and security, and his business is in database storage.
SnapStream’s Omekio ($600), which he demoed at the conference, uses database and network management open-source technologies to find sensitive data on packets and blocks. It doesn’t sell in volume, but the technology is clearly something people want to use. SnapStream has already been very successful, and is now trying to establish itself in the enterprise market, where it will face competition from Splunk (SPLK), Hadoop company MapR and hardware makers.
It is now a significant part of the market for secure network-side storage that Amazon (AMZN) is exploring as a building block for its Web services. These clouds are not exactly ideal for everyday use, but demand for them has been growing dramatically. If your organization could virtualize all the moving parts in its network, it would no longer be too far along in its migration.
There’s a lot of opportunity for startups to offer highly secure, high-performing alternative technologies. Just don’t confuse them with Google’s open-source container platform.
Daniel Salzberg is a senior editor for Cloud Foundry.