Okay, one of the most important things that I failed to consider and forecast was the time before the first sighting in the wild of a Borg picture and reference to assimilation. Too late now.
Clearly, even blogging policies can be a source of pain and evidence of malevolence when things are so confused and the feelings of loss are strong. However, when I look at what is being saved at Sun I have to feel positive about the idea that a company is going to take great technology and make money out of it. Even the switch in manufacturing, from build-to-stock to build-to-order, and simplification of the supply chain, reflects a refreshing willingness to take fundamental parts of the business and re-engineer them to improve performance.
So, before getting on to the accuracy (or otherwise!) of my predictions, let’s stop and consider the basics. Sun had a lot of technology, and a lot of IP, but some of was more valuable than the rest. The crown jewels were, and are, in my view, Solaris, SPARC, and Java. The application server, directory server, identity management suite, and so on, were and are all good (and as a company we use some of theme extensively), but they are largely “me too”. Oracle has committed to significant investment in Solaris, SPARC and Java, and that has reduced uncertainty considerably among users, as several Solaris and SPARC customers stated during the webcast. Further, there’s now a real sense of excitement and direction around the products, and a clear view of the ways in which Oracle will differentiate themselves in the marketplace as a server vendor.
Now to my predictions, and how I did. It’s fair to complain that I made them only hours before the webcast, but in balance I’m not sure that I had more insights at that point than several months ago. Also, I wasn’t overly specific in certain areas, which makes it easier to achieve a hit (or, at least, a near miss). Anyway….
On servers and storage, I think the summary I gave was reasonable, with an exit from commodity x86 (and Windows mentioned in particular) but a continued commitment to systems build on clusters of x86 servers running Solaris. There were a lot of mentions for Exadata, or the Oracle Database Appliance as it seems to be called now, and Sun Storage 7000. On processors, SPARC got the focus, and it was notable that the UltraSPARC T2 and T3 got top billing, along with a reasonable array of somewhat nebulous future versions which were intended to imply a dynamic future. The SPARC64 range, that results from the partnership with Fujitsu, got a mention, but its roadmap was largely incremental and lacking in excitement. I have to feel this was deliberate. On Java, as I thought, there was very little change.
Looking at what I got wrong, or missed, it was certainly the case that there were repeated reassuring noises regarding Linux, and the 4000 customers they have on support. I think I expected this, and it may be that Oracle realises that this is an important market for them. However, I still believe that the focus will gradually drift to Solaris, even if Linux remains a Tier 1 supported platform. Also, I missed the significance of the Netra range, built for the telco market. It was clear that Oracle regarded this as a significant opportunity. Finally, I wasn’t sure about the commitment to JavaFX, but it really does seem that they want to build more of their own products on this platform, and will invest to make it a competitive tool for RIA.
You have to wonder what Sun might have done had they made some of these changes before they were acquired, but I think it’s very hard to change things from within, no matter how senior you are. One of the most disappointing aspects of Sun’s performance in recent years has been the fact that they had great products and great technology, but didn’t seem able to deliver on many of the significant initiatives that they started, like the Sun Cloud.