Apple is expected to launch its iPhone 5 this week, and like all Apple events, this one is generating more than its share of speculation. It’s like a red carpet event for the Apple aficionados, but rather than focus on that, I’d prefer to look at the industry implications of what’s rumored to be happening.
First, the cloud. Amazon stole some considerable thunder from Apple with its Silk split-browser model because Amazon exploited a true cloud computing service to enhance a customer’s service. That’s the service layer model of the future. There are some indications that Apple will do more than just spin a “content hosting and sharing” model of the iCloud, including inter-iPhone messaging, but it’s not clear that these offerings are going to be positioned as part of the cloud at all, which I think would be a further mistake on Apple’s part. Amazon is a retailer, not a search giant, and thus a whole new kind of competition for Apple. They need to take these guys seriously.
Speaking of clouds, France Telecom has again said that the cloud is of critical importance to it and that it’s looking hard at the union of IT and networking needed to create it. This fits with our general survey results: Operators are increasingly seeing the cloud as the thing that hosts the service layer and the thing that hosts specific cloud-computing services. The question is how the two get combined.
Oracle may have a handle on it, and in its OpenWorld event this week it’s expected to launch a “big data” Hadoop-based cloud-distributable data engine. Hadoop is becoming an attractive model for at least some applications of cloud-distributed data, and it’s interesting to speculate that it might even become a standard strategy for hybridization. Oracle is also launching a NoSQL database, and one question here will be the extent to which Hadoop-like distributability is limited to NoSQL models. Most enterprises are, in my view, way too dependent on SQL to dismiss it even for cloud distribution.
Juniper has its own cloud announcement today, Junosphere Lab. The company launched its Junosphere notion with a classroom strategy that I think tossed away a pretty good concept on a weak first application. The lab app isn’t all that much better. Operators tell me that they really want a strong approach to multi-provider federation that they’re more in control of than the “umbrella” Open API Service of competitor Alcatel-Lucent. Junosphere could be a place where federation of service-layer (and even connection-layer) elements are hosted and even a place were operators host some service-layer features. Junosphere Lab steps closer to that mission but it’s not there yet, and we wonder if that’s because Juniper hasn’t advanced its own service-layer stuff far enough to support the broader mission the operators want. At any rate, Junosphere is a good idea if it’s used right, and Juniper, you need to stop dipping your toes into the applications and do something worthy of the concept—just like the service layer.