Original URL: https://www.theregister.com/2009/03/18/ibm_buying_sun/

IBM 'in talks' to buy Sun Microsystems

Schwartz dives for cover in Big Blue

By Chris Mellor

Posted in Channel, 18th March 2009 10:34 GMT

IBM has opened acquisition talks with Sun Microsystems, raising the prospect of a massive consolidation of the software, server and storage markets.

According to the Wall Street Journal IBM has mooted a price of $6.5bn. Sun is currently capitalised at $3.7bn ($4.97/share), but its share price has persistently fallen since the heady days of the dot com boom and has under-performed that of its computer systems competitors over the last few years, making it a much less expensive purchase now that it would have been three or four years ago. Legions of long-term Sun investors who have seen the value of their Sun holdings decline drastically will breathe a huge sigh of relief as they see the potential to make some money at least.

Any deal would likely bring very close scrutiny from regulatory authorities, given both firm's roles in the server, storage and systems software markets.

The Journal goes on to claim Sun has been hawking itself around the industry recently, looking to be bought, with HP also mentioned as one potential sugar daddy. If true, this confirms the perception that Sun has been impoverished by CEO Jonathan Schwartz's strategy of moving into open source software and relying for revenue growth on converting Solaris, Java and MySQL developers' into purchasers of Sun's servers, storage and services.

Unfortunately for Sun execs and investors, the free software download conversion rate into revenue dollars has been derisory, with Schwartz having promised soon-come blue revenue skies for what seems like years. Sun has continued to make money but its revenues have under-performed compared to its peers, and the company has been forced to go through rounds of restructuring and staff cuts - the most recent in January - to try and keep costs in balance with revenues.

Activist investor Southeastern Asset Management became Sun's biggest shareholder recently and some of the recent restructuring efforts have been thought to be influenced by this. Schwartz has consistently said he wants to work with this investor to increase shareholder value.

Selling Sun to IBM would give Southeastern Asset Management a profit on its investment and could be what it has been working for all along, and why it ponied up investment dollars in the first place.

There has been widespread industry speculation about Sun's future as analysts and commentators wonder how sustainable has been the board's and investors' faith in the Schwartz-led company and its strategy. But the fit of Sun's vast portfolio of assets into potential purchaser's businesses has always been a problem.

So why would IBM want a company which, at first glance, fits awkwardly into its portfolio of systems and services?

Sun product overview

Sun makes proprietary Sparc RISC servers, x86 servers and has storage and networking hardware, its Open Solaris operating system and allied software, its recently-acquired MySQL open source data base business, the massively popular Java software line, and its services business.

It has its own server chip design business and has persisted with this, in conjunction with Fujitsu, devising a line of multi-core processors to try and keep its servers ahead of the commodity x86 price/performance curve. IBM has its own RISC PowerPC chips which overlaps with Sparc. It's possible that the Sparc chip business will be offloaded to Sun partner Fujitsu, with the Sparc customer base being transitioned to PowerPC or x86 systems.

Sun's servers will overlap with IBM's PowerPc servers, and their users would surely be transitioned to IBM's own kit with the Sparc server line heading to the proprietary server cemetery. This transition will be eased by Solaris running on both Sun and IBM servers.

Solaris - Sun's version of Unix - is in terms of uptake a massive Sun success, with both HP and IBM supporting Solaris on their x86 servers. It is more popular than IBM's own Unix version, AIX, and there would the opportunity to merge the two - SolarAIX anybody? - and save development and support costs. How IBM would deal with the open sourcing of Solaris compared to AIX is problematical. It may be that it would end up following Sun's lead and open-sourcing a merged Solaris-AIX product.

There are also IBM's Linux activities to bring into the equation, but open source Linux and open source Solaris are not that far apart to create any insuperable difficulties, and an acquisition could considerably strengthen IBM's open source credentials.

In storage Sun would bring its Storage Tek mainframe tape business offering IBM an immediate consolidation opportunity, making it a virtual monopolist in the supply of tape formats and libraries to mainframe customers. The StorageTek disk array line would overlap with IBM's own DS4000, 5000, 6000 and 8000 array products and require some heavy migration work to remove overlaps and cut development and support costs.

IBM is not doing too well with its disk storage arrays, and a StorageTek array line and customer base infusion could be a good idea. There is scope for consolidation of the two companies' deduplication efforts, with IBM's Diligent line probably emerging triumphant and Sun's deduplication supplier FalconStor losing out.

Sun also has its Open Storage 7000 products with no RAID controllers and a comprehensive and sophisticated software storage stack, including its ZFS file software, running in a server. This is intended to undercut proprietary storage arrays, like IBM's DSX000 line and Sun's own StorageTek arrays and IBM may decide that its a storage revenue deconstruction exercise it doesn't want to pursue.

On the other hand it may conclude that undercutting EMC, NetApp and other proprietary storage array revenues with an Open Storage strategy could reap benefits, particularly with EMC, via VMWare, being so close to new server rival Cisco. Hacking away at EMC revenues could be an agreeable prospect for IBM strategic planners.

Software

The same would be true for Sun's upcoming Open Networking line where Sun is similarly intent on using software and commodity hardware to undercut and out-innovate proprietary switches and routers. That would be a potentially attractive way for IBM to get back into the networking business and chip away at Cisco. Sun's existing networking gear is not competing with any existing IBM products, and represents an expansion for its product range.

Sun will bring a very large open source product line to IBM indeed, with the crown jewels in terms of popularity being MySQL and Java. IBM has its own DB2 product line and MySQL would compete with that - although one would think Big Blue's marketeers could separate the two fairly easily. MySQL might be targeted at SQL, for example. Big Blue might also make better revenue-generating use of Java.

The future of Open Office is possibly good under an IBM stewardship as it gives Armonk the opportunity to take on Redmond in desktop software - the gut appeal is instant.

It's become a truism that Schwartz has failed to protect Sun's proprietary high-end server business, failed to capitalise on the StorageTek product line of tape libraries and disk arrays and, the most damaging, failed to monetise Sun's vast numbers of free software downloaders. Will IBM do any better?

We could see how IBM could move around Sun lego blocks, merging some with its own, discarding a few, and coming up with a resulting product line-up that better equips it to meet and beat HP; respond to Cisco; have another go at Microsoft; gain revenue from Sun's vast download base; get a mainframe tape monopoly... and sell its own products into Sun's large customer base.

It would be a hugely consolidating move and better enable IBM to compete with Mark Hurd's HP which has emerged as its strongest competitor in years. The cost would be larger than the $5bn Cognos acquisition, but the potential benefits are larger too.

If an acquisition does happen it's likely that Schwartz will leave Sun, possibly along with many of his senior execs. He will, in the minds of many people, have failed at Sun.

The picture of IBM swallowing Sun will be, like that of HP gulping Compaq which in turn had eaten Digital Equipment, an iconic burial moment for the minicomputer and workstation boom of the Eighties. Only HP is now left of the band of silicon revolutionaries who set out to dethrone the mainframe - Apollo, Data General, DEC, Wang and the others who are now all gone. Mainframe survivor IBM absorbing Sun is an apt confirmation of its status as the survivor of the IT age.

IBM has issued a 'no comment' about any talks with Sun, which in turn has said: “As a matter of policy, we do not comment on market rumours about our Company.” ®