Beachhead for on-demand computing

Jonathan Swartz, COO at Sun has been discussing Sun's on-demand grid computing offerings for some time now. Here's how he proposes Sun would pitch it to customers - (Quoted) "..we'll tell our customers - if you have comptutational workloads, that require 10's or 100's or 1,000's of cpu's, for defined periods of time (ie, 5 hours, or 3 days or 3 months) - discrete jobs like rendering a movie, or doing a monte carlo or geophysical simulation, or modeling a protein - then we can run your loads on demand for less than anyone in the industry."...

Sounds like Sun's beach-head for the CPU grid will be CPU intensive 'peak workload' computing requirements. I think this also gives a peek into Sun's sense of where on-demand CPU grids are in terms of their market maturity and customer acceptance. Sounds like they're still a good while away from prime time. (I think storage grids are well further than that though).

While Sun goes about developing its on-demand infrastructure business, its also worth noting how their approach contrasts with that of vendors who've seen success in the on-demand applications space. The success of ASP's like Salesforce.com, the erstwhile Upshot, NetSuite, came from establishing beachheads in the small-midsized business (SMB) segment, where customer needs are ideally served via a utility style commodity model. That essentially holds true for eBay and Google as well.

Beyond their beachheads, on-demand ASP's have typically started to address more specific customer needs as well. They provide options - public API's, custom toolkits, custom deployment options, professional services - to address more sophisticated pockets of demand. Always, the basic commodity service has utility style pricing based on a standard unit of measure, and value layered on top of the basic service results in pricing and costs that are more and more specific. In serving this segment, the utility model actually becomes weaker.

Sun seems to be looking to niche pockets of demand for a beachhead. Or quite likely, they don't think they need a beachhead, and can afford to slowly evolve from their current position. In which case, their grid story simply serves to capture mindshare by projecting a certain thought leadership. This is unlike the situation that ASP's faced when they started out.

My take - I would'nt expect major hits from their on-demand CPU grid if its primary business case is not based on serving high volume, low price markets. Its also possible that direct uptake in SMB markets will be challenging because Sun technology has'nt traditionally been aligned with the SMB market. SMB exposure is more likely via indirect channels, through uptake of infrastructure grids among ASP's themselves.

Coming sooner than you thought - On-Demand Storage Grids

Industry pundits out there have been quick to trash Sun Microsystem's deal today to buy StorageTek. That leaves me scratching my head. Putting aside the number crunch - deal price, earnings accretion, cash outflow, etc. The numbers actually work out just fine, going by that sort of tactical financial analysis. But I'm not sure why the bigger market picture is'nt clearer to more analysts out here. Sun's got the R&D discipline, and an interest in the storage market. And StorageTek is a gorrilla, carrying technology flab and looking for a workout. Put the two together, and there's a strong chance that Sun can make real good out of this.

Sun has recently been pitching the vision of storage as a utility service - its storage grid - just plug in for $1/GB/month (OK, not as simple as 'just plug in', but bear with a little market-speak here). With StorageTek's customer base, sales strength and storage expertise, Sun gets to jumpstart this innovative service model. I'd predict with these two companies getting together, Sun can significantly accelerate adoption of storage grid services among businesses with big storage needs. And to be sure, storage does quite perfectly fit the description of a computing commodity, ripe for the utility model. Even before adoption of CPU grids starts to happen. Not to forget that data storage needs are growing explosively - 70% annually by some counts. Storage services offer a low risk entry point for customers to start exploring the reliability, security, efficacy and cost savings from grids. And businesses can start by using them in baby steps, incrementally growing in maturity, first with secondary backups, to archiving, to primary backups and finally to primary storage.

And while we get closer to that scenario becoming more of a reality, in the immediate future, Sun now has the opportunity to pitch StorageTek customers with computing, storage and network solutions all under one roof. I'm sure that will help Sun snag some share from IBM, HP and Dell; Sun surely needs that. But I'll be waiting for those new customers to cosy up a bit with Sun, and then...hallelujah, give them the gospel of the grid and plug them right in!

Its classic acquisition synergy. StorageTek has the experience and relationships. Sun has the smarts and vision. Together they can build real value in the storage marketspace. So its worth repeating - why are'nt more analysts seeing it this way?