On December 2, Amazon announced a change to its EC2 reserved instance pricing model. Reserved instances allow customers to reserve computing power for three years in contrast to the on-demand, pay as you go model of cloud computing. Reserving instances for a fixed duration entitles customers to deep discounts in comparison to hourly pricing. Previously, Amazon Web Services customers were required to pay up front, in full, for either one to three years of reserved instance usage. Last week, however, Amazon unveiled two additional pricing options for reserved instances marked by the choice to pay either a partial amount of the reserved fee up front or no up front costs at all. Customers who pay no up front costs benefit from the cash flow advantages specific to spreading out their payments over the duration of the reserved instances term. However, paying up front for a three year reserved instances term gives customers a 63% discount in comparison to on demand pricing over three years, whereas the no up front payment delivers lesser savings, on the order of 30%.
Amazon’s reduction of its reserved instances pricing was announced days before a price reduction on outbound data transfer fees ranging from 6% to 43%, depending on the region and the volume of data transferred each month. Meanwhile, Amazon also recently announced an API function, PutRecords, that streamlines the insertion of data into a stream of Amazon Kinesis data. The PutRecords function can transmit as much as 4.5 MB of data into an Amazon Kinesis stream by means of a function that sends a maximum of 500 records, each one as large as 50 KB, into a Kinesis stream via a single HTTP call. The bottom line is that Amazon continues to innovate and remain competitive amidst growing competition from Google Compute Platform and other IaaS vendors whose products, pricing models and partner ecosystems are increasingly maturing and posing more of a competitive threat to Amazon’s market share dominance in the IaaS space.