Price-focused v. Value-based IaaS Vendors

Written by: Louis Corso

Amazon’s new auction based pricing strategy brings to light some philosophical differences between the price-focused and value-based cloud vendors. While they are ideal for non-mission critical tasks such as software development, price-focused vendors are less prepared to meet the needs of a growing demand for enterprise grade performance and reliability.

Several blogs and industry publications have recently been posting stories about the new auction style pricing model Amazon has adopted for their unused capacity in EC2. Here’s the rundown of how it works:

Users provide a maximum amount they are willing to pay per instance for a particular EC2 instance size. This “bid” is the compared against Amazon’s spot price for that instance, which is determined based upon current demand for and supply of that instance. If your bid is above this price, you get to use the instance until the spot price exceeds your bid. The spot price will fluctuate throughout the day, and you instance will be frozen and unfrozen as the price is adjusted above and below your bid. These “Spot Instances” can be cancelled at any time once the user no longer needs them.

For Amazon, this is a great idea because they are selling off resources that would go unused under normal conditions, moving them ever closer to full utilization of their infrastructure. For calculation based projects, like media processing or financial analysis, this may work out great for the users as well because they end up paying a lower per-hour rate, assuming that they can tolerate indeterminate completion times. This seems to be a win for both parties.

Pricing has been a big subject here at ENKI recently as we’ve started to notice a rift which is forming in the Infrastructure-as-a-Service market. On one side of the canyon are Amazon and the other price-focused vendors. On the opposite side are the value-based providers, such as ENKI and others, who are focused on service and service-level agreements.

True enterprise level uptime, 99.95% or better, has an actual cost in hardware and labor which will never be covered by the $0.085 per hour Amazon charges for a small Linux instance. In particular, Amazon’s solution has a very hard time guaranteeing application uptime because there are no partners, to my knowledge, who provide application level monitoring and operations services to the Amazon cloud. However, there are many customers for which application, and even hardware uptime guarantees aren’t as critical. In particular, software development, which does not require strict adherence to SLAs as long as the data remains intact, may be willing to accept the compromises necessary to achieve Amazon’s lower prices.

With the pricing strategies they’ve put in place, including low cost CPU hours and the auctioning of unused resources, Amazon will by necessity start moving away from having guaranteed SLAs. This “race to the bottom” is at odds with the interests of a growing number of enterprises who are starting to explore the potentials of outsourced IT. The needs of this new customer base will require service levels that can only be met by vendors committed to a high level of quality and services. Value-based providers are well equipped to take on this new generation of more sophisticated IaaS consumers.

Set as favorite

Bookmark

Email this

Trackback(0)

TrackBack URI for this entry

Comments (0) add comment

Write comment