Sun could be about to offload its failing 6920 external storage drive array to Hitachi Data Systems, or so the rumour-mongers say.

But is it or isn't it? The weakening silence is laden with pregnant pauses, heavy hints and dogs that are not barking. HDS, from the land of the rising sun, says nothing. Sun, from the land of the wished-for rising sun, is not saying "no" and not saying "yes" either.

Instead it states "Sun is continuously investigating ways to reduce costs and manage our product portfolio. We are also in regular discussions with our partners about how to better serve our customers. At this time, we are not prepared to disclose any details around changes in our product roadmap. As a company policy, we do not respond to market rumours."

The rumours relate to its mid-range, virtualising, external drive array control, the 6920. This uses technology acquired when Sun bought Pirus for $160 million in 2002. Sun also has an OEM arrangement with Hitachi Data Systems (HDS) for its high-end TagmaStore drive arrays with their intelligent, virtualising, external array controller, the USP. These are Sun's 9990 products.

This has been implemented in mid-range form as the NSC55 (Network Storage Controller). Obviously it competes with the 6920 in the mid-range market for the management and virtualisation of multi-vendor arrays. HDS has recently announced it has sold 4,500 virtualising array controllers. Sun has never announced a 6920 sale; one of those deafening silences perhaps.

There have been rumours for some time that the 6920 business is up for sale. The latest stories suggest that the deal may have been done.

A difficulty is in understanding the logic. HDS already has mid-range virtualising external array controller with its NSC55. Why should it want to buy another?

To incorporate newer and/or better technology? Perhaps. The 6920's interface has been praised and perhaps its partitioning functions are well-regarded. But why buy a product line when IP (intellectual property) licensing would do?

The logic of Sun letting the 6920 go can be seen. It would save Sun engineering expense, simplify its storage product line, replace the probably ill-selling 6920 with the NSC55 and its momentum, and remove a marketing message inconsistency concerning two different virtualising array controller product technologies.

The difficulty for external observers is the HDS motivation. Perhaps one has to look outside the virtualising array controller space. HDS has its new BlueArc NAS products to sell, also a virtual tape library and content-addressable storage. It could be that, in exchange for taking on the 6920 and migrating any customers to the NSC55, HDS will obtain a deeper relationship with Sun, meaning Sun selling additional HDS products.

Both Sun and HDS understand that drive arrays are a fast commoditising business. The higher gross margin money comes from intelligent storage management functions. Sun's technological interest is in building so-called hybrid storage products combining heavy processing capabilities with the disk platters, Honeycomb and Thumper for example, also in developing ZFS.

If HDS and Sun have similar views on how to grow revenues from building intelligent systems situated at the server/storage interface, and their product development strategies are in synchrony then a closer relationship would be good news for both parties.