The Case for Matching Sun & EMC
Should Sun be saved from the ignominy of being sold for parts? Might EMC be the company to do that?
Our world is never static, but as I write this, we are going through a particularly nasty period of upheaval.
Print publications are dying left and right. The latest examples are US News & World Report and The Christian Science Monitor, which have thrown in the towel on their print editions. Forbes magazine just merged its print and online divisions, resulting in 43 more journalists looking for work. Happy holidays.
Likewise, the tech industry is working through some wrenching change, some more than others in the difficult economy.
The Problem With Its Parts
One of the problems with Sun's continued slide is that its UltraSPARC line of chips is falling out of favor and there's little the company can do to reverse this. The competing x86 architecture has found new life with the successful jump to 64-bit computing systems, and enterprises tend to stay with a commodity platform such as x86 over one built by a single vendor.
Think about this: Intel is crowing about its new chip that does two threads per each of its four cores, while Sun has a chip that does eight threads per core and has eight cores. The benchmarks for its new T5440 are incredible. Intel can't come close.
But perception is everything. Sun may make good hardware but it's an x86 world, even if x86 servers can't match UltraSPARC for performance. Meanwhile, IBM is holding onto a nice piece of the high-end server pie with its Power line, and HP has its Itanium products; there is just only so much room for non-x86 platforms. Quarter after quarter, Gartner and IDC report server stats that show that Sun is selling hardware to a shrinking customer base, which is migrating to IBM or HP.
Since Sun's most recent quarterly loss $1.7 billion, speculation is ramping up about whether parts of the company may be sold off to save it.
Industry watchers doubt that board members of Sun would let the company be carved up like a Thanksgiving day turkey. They may have no choice.
Reuters reported that Southeastern Asset Management, which holds 20 percent of Sun's shares, might go around Sun's board of directors and talk to "third parties" about alternatives.
Sizing Up the Prospects
Fujitsu's name gets thrown around the most as a potential buyer because of the lengthy partnership between the two firms.
IBM? It really doesn't need Sun and is doing fine converting Sun customers to its hardware as fast as it can. Buying Sun would net it Java, but IBM has nearly taken it over already.
As for HP, that firm is transitioning off its RISC chip, the PA-RISC. All those engineers are now with Intel, working on the Itanium server lines and competing with UltraSPARC. Dell doesn't appear suited for big acquisitions right now, unless the betting is on an SGI purchase.
So how about EMC? The company, considered the first name in storage with an aggressive acquisition history, appears to be a partial match.
EMC specializes in network-attached storage (NAS) and storage area networks (SAN), while Sun's StorageTek line specializes in tape-based storage. Granted, EMC doesn't like tape storage, but it can't ignore the market for tape storage, either.
Also, Sun possess plenty of other technologies that have been brought to bear in the storage market. The Sun Fire X4500 server, a.k.a. Thumper, combines a Sun server running Opteron chips, the Solaris operating system and ZFS file system.
Server performance is becoming an issue, as Oracle pointed out with its HP-built Exadata database servers. A plain old server just doesn't work any more. Servers need a little software and hardware assistance to move data at a faster rate. So vendors are now building accelerated servers to improve throughput. This could be a compelling combination for EMC to ponder in Sun.
With its own technologies combined with the Niagara 2 chip, Solaris and ZFS, EMC could come back with a product line that could challenge the HP's systems. Sun technologies could also augment the new EMC line of cloud storage systems called Atmos, as those clearly could use some kind of intelligence and computing power.
Then there's the software side. The potential of EMC and Java is huge, given EMC's efforts to move into a more software-oriented strategy -- from RSA security to database replication, content management, backup and recovery. Java's server-side strengths could fit this strategy well.
Then there is the potential match with VMware, such as putting the Java Virtual Machine in the same layer as the virtual machine layer, which leads to the possibility of controlling Java apps across all the virtualized systems from a single layer and single VM; thus, the virtual server knows what all of the apps are up to.
Databases are rarely virtualized, but with MySQL in-house, EMC could find a way to make at least one database much more suitable to virtualization. The competitive advantage there would again put EMC in a unique position.
There is one big snag to this EMC/Sun theory. EMC has displayed no interest in the open source movement, while Sun has embraced it in a big way.
This could prove a no-win situation if EMC picks up Sun and open source advocates demand that it open source other products -- especially if GPL-licensed products start mixing with EMC software. If EMC balks, it could wreck the goodwill Sun has built over the last few years. Then again, goodwill doesn't pay the bills.
Plus, the cultures between the two could lead to some real problems. Joe Tucci is as no-nonsense as they come, and his gruff nature permeates the staff of EMC. That's led to clashes with VMware, and could cause problems with the laid back culture of Sun.
But looking up and down the EMC product lines, it's clear they offer the best chance for Sun's products to not only live on, but to thrive and offer customers more innovative choices.
Andy Patrizio is a senior editor for InternetNews.com based in the San Francisco bureau.
Article courtesy of InternetNews.com