In-Depth
With Sun Deal Behind It, Oracle Goes Retro
Oracle's post-Sun announcements harkened back to the heyday of single-stack computing: the IBM Corp. of the 1960s.
After nine months in limbo, Oracle Corp.'s acquisition of the former Sun Microsystems Inc. is now complete.
Although Oracle still has much work to do, and though questions persist about the viability of some of Sun's most cherished assets (chief among them its flagship UltraSPARC chip business), the computing giant offered most of the right answers, industry watchers acknowledge.
In fact, notes Gordon Haff, a senior IT advisor with consultancy Illuminata Inc., Oracle's post-Sun announcements recalled the heyday of single-stack computing: the IBM Corp. of the 1960s.
"[The surprise] was the strength of the all-Oracle stack message. Oracle CEO Larry Ellison, among others, made it clear that the IBM of the 1960s was their integration model," writes Haff on his Illuminata blog. "There were a few references to selling 'best of breed components' to customers who wanted to purchase that way. But promoting the benefits of buying a complete hardware and software stack designed to work together was one of the other overriding themes of the day."
One dead giveaway, notes Haff, was the complete absence of a once-de-rigueur industry buzzword: "I'm not sure I heard 'heterogeneous' -- one of the terms that computer companies like to use even when they don't really mean it -- uttered once during the five hour broadcast"
Long, Strange Trip
Oracle Corp. does have a knack for follow-through, particularly when it comes to its long-and-drawn-out acquisition. Its pursuit of the former PeopleSoft Inc., for example, is the stuff of M&A legend, and -- although almost all of Oracle's subsequent acquisition moves were considerably less drawn out -- its April, 2009 bid for Sun proved to be an obvious (and glaring) exception.
Few people seemed to doubt that Oracle would eventually take ownership of Sun. At issue was the condition of Sun that Oracle would take over. Fear, uncertainty, and doubt (FUD) necessarily attend any acquisition. Even when a pending takeover move positions a player to dominate a market, competitors invariably exploit the quiet time between the announcement and closing of an acquisition to promote FUD loudly.
This was doubly so in the case of Sun, which had been publicly shopping itself around --first to IBM Corp., but also to Hewlett-Packard Co. (HP) -- for weeks prior to Oracle's move. Sun started 2009 in a comparatively weak position: it went through two layoff cycles in 2008, and by January of last year the company was perceived as taking on water, if not in imminent danger of foundering. When a healthy (and, in Oracle's case, vigorous) company buys less-healthy competitor, it can be perceived as a white-knight rescue move -- the former Hyperion Solutions Corp.'s acquisition of the former Brio Software Corp. is such a textbook case. Although Oracle's bid for Sun surprised and confounded many industry watchers, others saw potential complementary areas -- even synergy -- in the pairing.
The question for many in the industry concerned what shape Sun would be in once Oracle finally took possession.
As the regulatory process dragged on, and as both HP and IBM continued to nibble away at Sun's market share, the erstwhile Unix champion's position continued to erode. The server market as a whole was substantially down in 2009, but -- as market watchers Gartner Inc. and IDC noted -- Unix stalwart Sun suffered disproportionately. In Q4 of last year, for example, Sun actually led the Unix market in terms of unit shipments, even as it fell to third place in terms of total revenue. Sun was slashing its prices to keep pace with FUD-mongering rivals HP and IBM.
Harbinger of Things to Come
Sun and then-acquirer-to-be Oracle weren't completely defenseless.
The duo struck back starting in August, first running a series of advertisements touting the combination of Oracle software and Sun hardware (in which Oracle/Sun took direct aim at transaction processing champ IBM), and then, in September, introducing a new Sun-powered Oracle database server, Exadata Version 2, the successor to the Oracle Database Machine (a super-charged data warehouse system) that Oracle and then-partner Hewlett-Packard had unveiled in September of 2008.
At the time, many believed that Oracle chief Larry Ellison was particularly piqued by former partner HP's cannibalization of Sun's server business. When introducing its Exadata v2, Oracle announced plans to phase out HP-based Exadata 1.0 systems. Illuminata's Haff, for his part, cites the introduction of Exadata v2 as a kind of exemplar demonstration of Oracle's single-stack pitch. "With Sun and Oracle's announcement of a database appliance last fall, there could no longer be any doubt that delivering factory-integrated stacks from server to storage to software was a big part of this acquisition," Haff wrote in the same blog post.
Curt Monash, a principal with database management and analytic database consultancy DBMS2, reports that Oracle officials believe a big chunk of enterprise customers will eventually standardize on Exadata v2 as a platform for their Oracle database workloads. Unlike its HP-powered predecessor, the Oracle/Sun Exadata v2 is positioned as a platform for both OLTP and analytic (or data warehousing) workloads. “[Oracle] believes the ‘bulk’ of [its] business will move over to Exadata-like technology over the next 5-10 years. Numbers-wise, this seems to be based more on Exadata being a platform for consolidating an enterprise’s many Oracle databases than it is on Exadata running a few Especially Big Honking Database management tasks,” Monash writes on his DBMS2 blog.
Although Monash spoke with Oracle’s Juan Loaiza ahead of Oracle's lengthy Oracle/Sun broadcast, Loaiza was clearly on message. The vision he outlines is of an Oracle that -- like IBM, particularly with its System z -- clearly plans to protect and promote its best-of-breed crown jewels.
In at least one respect, Monash observes, that means Exadata. “Oracle doesn’t make its storage-tier software available to run on anything than Oracle-designed boxes. At the moment, that means Exadata Versions 1 and 2,” he writes. “Since Exadata is by far Oracle’s best DBMS offering -- at least in theory -- that means Oracle’s best database offering only runs on specific Oracle-sold hardware platforms.”
The Single-Stack Pitch
Concomitant with Oracle’s single-stack pitch, Haff highlights what he sees as another Right Answer. "The other big theme could be summed up as something along the lines of: Sun had great innovation but executed really poorly," he points out, citing Ellison's remark that Oracle "just need[s] to do a better job of taking engineering output and delivering it to customers."
Industry veteran Charles King, a principal with consultancy Pund-IT, has a similarly sanguine take on Oracle's post-Sun situation.
First, he notes, Oracle's single-stack pitch actually makes a lot of sense -- and, to a degree, is already practicable -- thanks to the historically close relationship between it and Sun. "On the plus side, the pair's long, close association should aid their integration and help drive new and future synergies and product development. In other words, the pairing looks good technologically," he notes, conceding that "Oracle might also be a better fit culturally for Sun than either HP or IBM."
King wasn't quite reassured by the enthusiastic, pro-UltraSPARC rhetoric adopted by Ellison and other Oracle principals, however. Perhaps more than any other chip architecture from any other vendor, UltraSPARC is closely bound up with Sun's identity; at the very least, King notes, Oracle had to commit to its importance. Anything less could have led to revolt on the part of long-standing customers, many of which run the Oracle database on top of UltraSPARC hardware.
"Ellison's enthusiasm for Sun hardware should be considered in the context of client management. Oracle is the database of choice in a large majority of Sun-populated data centers. The last thing Oracle wants is for Sun customers to believe that UltraSPARC is on a short road to nowhere," he points out. "In the short term, falling faith in Sun hardware would impact Oracle's bottom line -- not to mention undercutting Ellison's claims that Sun assets will become immediately profitable."
Haff, a seasoned industry veteran (and one with a special interest in and knowledge of microprocessor architectures), is likewise dubious.
"One technology area where many of us have had trouble imagining Oracle making large investments is processors. Sun chips [have] lagged in recent years and, even with major new talent and cash infusions, development cycles are relatively long," he points out. "Nonetheless, Oracle presented a road map going out four processor refreshes in the UltraSPARC Tx -- originally code-named "Niagara" -- and Ellison suggested that Oracle could do 'clever things in the chip like encryption and compression' and 'a lot of things we used to do in software can find their way into the silicon.'"
This is far from tendentious or cutting-edge, Haff notes: IBM has been doing similar things in its POWER-based CMOSes (on both the mainframe and System p/System I sides) for quite some time, so it isn't inconceivable. That said, he suggests, Ellison and Oracle were perhaps less believable in the UltraSPARC than in other areas.
"I still consider it an open question whether this enthusiasm for developing processors extends beyond a near- to mid-term desire to keep the Sun installed base happy and thereby buying stuff from Oracle, but at least for now UltraSPARC has a future."