it en



Will Google IaaS offering add fuel to cloud's price war fire?

IntroWhen asked where Google fits into the pantheon of big cloud players, panel experts at Cloud Connect Santa Clara in March all agreed it could be a big player -- if it wanted to. But what does that mean for the cloud vendor market and, even more important, for cloud consumers? Focusing on Google's Infrastructure as a Service (IaaS) offering, the Internet giant certainly has large-scale datacenters from which it could carve out capacity for Google Compute Engine (GCE). However, the company likely will face significant challenges in doing this. The publically available information about its architecture shows that it was purpose-built for the primary business, Internet search. Google has been able to extend this to other applications, such as Gmail, that can leverage the infrastructure. The GCE preview currently has two flavors of Linux, with no mention of Windows Server support. This could be an issue, as a generic IaaS needs to support a breadth of operating system (OS) images to outpace the current cloud market leader, Amazon Web Services. It is also unclear how well the Google automation infrastructure can support IaaS's more generic requirements. Windows Azure Platform as a Service (PaaS) required significant changes to enable the full set of features, such as a local domain name system (DNS), for a broader service. Microsoft announced that its now branded Windows Azure Infrastructure Services missed its very aggressive general availability (GA) goal by eight months, which indicates the amount of work required to repurpose from one cloud service model to another. And if we view the service layers as somewhat of a continuum rather than discreet, Google's current offerings would be higher in the PaaS and Software as a Service (SaaS) realm. Windows Azure would be closer to the IaaS boundary, which made Microsoft's job the easier of the two. GCE should be in GA in the not-too-distant future -- at least if the technological challenges are the gating factors.

What does Google IaaS mean for the industry?

But this discussion about Google's IaaS cloud offering really comes down to money. In the already saturated cloud market where the leaders must fight for position in a price-per-unit war, who will benefit from the entrance of another big player? Generally, a price war is good for the market; consumers get lower prices on whatever product is viewed as a commodity. One aspect of a commodity is that goods are indistinguishable between suppliers, allowing the customers to freely switch based on the best price. Currently, IaaS does not meet that criteria, so consumers see vendors of value-added products cratering their prices to grab market share in the hopes of being the last man standing. Some analyses hint at an endgame that leaves more than one player standing, but there is historical precedent for consolidation to a monopoly being bad for the market because that would cause prices to increase. And because of those historical situations, there are laws and regulatory bodies that might not permit the prices to rise to a profitable level -- leaving that single provider permanently "in the red." The three major players in cloud -- Amazon, Microsoft and Google -- have other "cash cow" businesses from which they can fund a significant amount of red ink for quite some time. Some onlookers have postulated that service levels will start to drop after price equilibrium is reached in an attempt to stem the losses. There are, however, other more likely scenarios. There is a desire on the part of the market to avoid vendor lock-in; IaaS is considered the path to this nirvana. This is not currently realizable as the three major public IaaS vendors do not currently support deployment of images from their competition. Once consumers choose a vendor, they are effectively locked in and pricing becomes more flexible (from the vendor perspective). Vendors will make a strong move to entice customers to adopt their value-added services beyond the commodity layer. And these services will have better profit margins and will increase the lock-in again resulting in price flexibility for the vendor. If neither of the preceding methods works, any or all vendors may not find it beneficial to continue participating. If all but one of the survivors exits, the survivor is free to raise prices. And the former customers of the exiting vendors will see their economics change dramatically. Price wars are very exciting from the customer side in their early stages, but such battles do not ultimately lead to a healthy market. We should be very careful in how loudly we applaud this street fight. It's more important to listen to market conversations on how to achieve sustainable economic benefits from the real value of cloud computing that go beyond lower costs.
<< Back