This week, Amazon Web Services data center guru James Hamilton wrote a blog post responding to comments made by Oracle co-chief executive Mark Hurd last week.
As Fortune reported, Hurd discounted the idea that his company cannot compete in cloud services with Amazon, Google, and Microsoft because those three companies spend so much more on data center infrastructure. Those three cloud giants invested an estimated $31 billion on infrastructure last year while Oracle dedicated about $1.7 billion, according to figures assembled by The Wall Street Journal from public filings.
Hurd’s contention was that the company with the highest data center capital expenditures won’t necessarily win the cloud battles. By offering faster databases and faster hardware, Hurd argued that Oracle can do more with less.
Hamilton, an Amazon distinguished engineer, noted that Oracle isn’t the only company that can speed up databases.
“All major cloud providers have deep database investments but, ignoring that, extraordinary database performance won’t change most of the factors that force successful cloud providers to offer a large multi-national data center footprint to serve the world,” Hamilton wrote.
Hamilton, who draws huge crowds at the annual AWS Re:Invent show in Las Vegas, went on to detail the importance of super-fast networking connections both inside data centers and between data centers worldwide. Amazon (AMZN), Microsoft (MSFT), Google (GOOG), and—yes—Oracle are racing to build more data centers in all regions in part because the distance between a data center and the people using it adds latency (or delay) to transactions.
The most efficient number of data centers per region is one. There are some scaling gains in having a single, very large facility. But one facility will have some very serious and difficult-to-avoid full-facility fault modes like flood and, to a lesser extent, fire. It’s absolutely necessary to have two independent facilities per region, and it’s actually much more efficient and easy to manage with three.
Read the whole post for details, but it is interesting that Hamilton took the time to respond to comments from a company that AWS seems to see as a cloud also-ran.
Oracle’s business selling online applications in the same mold as the Salesforce model is growing fast. But its public cloud business—in which companies rent basic storage, computing, and networking resources—is tiny. In Oracle’s last fiscal quarter that “Infrastructure-as-a-Service” business logged $178 million in revenue. AWS, which derives much of its revenue from IaaS, reported more than $3.5 billion in sales for its most recent quarter.
Get Data Sheet, Fortune’s daily newsletter on the business of technology.
Also interesting: Since January, Oracle’s cloud effort has been led by senior vice president Don Johnson, a former AWS tech guru.
The rhetoric between the two companies has been pointed for some time. At AWS Re:Invent two years ago, AWS executive Andy Jassy said that every database customer he talks to is unhappy with their vendor. He named no names, but Oracle leads the market in databases. Jassy, who is now chief executive of AWS, reasserted those claims this week on CNBC.
Say what you will about the state of Oracle’s public cloud business. Its strategy—or at least Hurd’s fighting words about it—have gotten Amazon’s attention.