Amazon built its storage solutions by leaning heavily on core principles from distributed systems and database research.

Judge as of 2011. True: Amazon's storage systems in use by 2011 (e.g. Dynamo, S3, SimpleDB) were plainly built on established distributed-systems and database research principles and extended them. Evidence would include Amazon's own technical papers (such as the 2007 Dynamo paper citing consistent hashing, vector clocks, quorum and gossip protocols) or independent technical analysis. Mostly true: this holds for Amazon's main storage systems, with a minor detail not supported. Mixed: some Amazon storage systems clearly drew on such research and others were mostly built without it. Mostly false: the link to the research is weak or limited to one minor system. False: evidence shows Amazon's storage systems were developed without relying on that research.

Set before the research began.

1. E1 Amazon's Dynamo All Things Distributed (Werner Vogels' blog) · 2 October 2007 Own record

   “Dynamo uses a synthesis of well known techniques to achieve scalability and availability: Data is partitioned and replicated using consistent hashing [10], and consistency is facilitated by object versioning [12].”

2. E2 Amazon's Dynamo All Things Distributed (Werner Vogels' blog) · 2 October 2007 Own record

   “We submitted the technology for publication in SOSP because many of the techniques used in Dynamo originate in the operating systems and distributed systems research of the past years; DHTs, consistent hashing, versioning, vector clocks, quorum, anti-entropy based recovery, etc.”

3. E3 Dynamo: Amazon's Highly Available Key-value Store (2007) mwhittaker.github.io (personal paper summary)

   “Dynamo is Amazon's distributed higly-available eventually consistent zero-hop distributed hash table (a.k.a. key-value store) that uses consistent hashing, vector clocks, quorums, gossip, and more.”

4. E4 Amazon's Dynamo All Things Distributed (Werner Vogels' blog) · 2 October 2007 Own record

   “To meet these stringent latency requirements, it was imperative for us to avoid routing requests through multiple nodes (which is the typical design adopted by several distributed hash table systems such as Chord and Pastry).”

Amazon's 2007 Dynamo paper says the system is "a synthesis of well known techniques", naming consistent hashing and object versioning (E1). Werner Vogels' blog post introducing the paper says many of Dynamo's techniques "originate in the operating systems and distributed systems research", including DHTs, vector clocks, quorum and anti-entropy (E2). An independent summary confirms that Dynamo uses consistent hashing, vector clocks, quorums and gossip (E3). The paper also describes a deliberate change from the multi-hop routing used in research systems like Chord and Pastry (E4), which fits the claim that Amazon "invented from there." However, the evidence covers only Dynamo, not S3 or SimpleDB, so it supports the claim for one main system rather than for all of Amazon's storage systems.