NEW YORK -- To succeed in the cloud, companies must change the way they consider their cloud database options, according to Amazon's Rahul Pathak, senior product manager for Amazon Web Services'
That's what Pathak told a roomful of cloud developers and architects at the 2013 AWS Summit in New York this week. "People tend to focus on database technology when choosing databases," he said. "What they should really be thinking about are the application's requirements, where they want to spend their time, and how it all matters to the business."
You can have the best of both worlds: Use both SQL and NoSQL models in one application.
senior product manager, AWS database services
With the right focus, Pathak said, organizations can ensure that they'll choose the best database for any cloud project -- a task that gets more challenging as database options become more numerous.
Stamford, Conn.-based Gartner Inc. reports that the market for databases used for reporting and analyzing data is growing rapidly due to the rise of big data. More and more, companies are jumping into the database and data warehouse fray -- without a clear idea of how to choose the right one.
To that end, Pathak had this simple advice for conference attendees at the Javits Convention Center in Manhattan: One size does not fit all. "If your data model is relatively simple and your updates are straightforward, NoSQL databases can be a great option and they tend to scale better," he said. "Or, if you'd rather focus on your application, fully managed databases are a great way to go."
When it comes to choosing between relational and non-relational databases, Pathak advised that a company should determine which capabilities are absolutely necessary. For example, if an application requires complex queries, a relational database is a good fit. If not, then a non-relational database might work better.
Pathak also cautioned his listeners not to box themselves in with just one type of database. "You can have the best of both worlds: Use both SQL and NoSQL models in one application," he said. "It's possible to manage different tiers in different ways." For example, one AWS customer has a fully managed database but chose to have complete control over its front-end layer, he added.
Database choices for AWS: A rundown
The myriad of cloud database options -- from MySQL and Redis to Cassandra and MongoDB -- can be daunting. Pathak acknowledged the crowded database market. To address AWS customers, he offered an overview of AWS database options and their differentiating features:
Amazon Relational Database Service (RDS). A fully managed SQL database service, Amazon RDS offers an array of database engine choices. "The goal of RDS is to take over all the muck associated with database management -- things like migration, backup, recovery and patching," Pathak said. Notable RDS users include Sega, TweetDeck and Samsung.
Amazon ElastiCache. ElastiCache is a fully managed caching service, protocol-compliant with Memcached. Users can scale cache clusters with push-button ease, Pathak said. "It has an ultra-fast response time to read scaling," he noted.
Amazon Dynamo Database (DB). Dynamo DB is a fully managed NoSQL database service. "It can store and retrieve any amount of data," Pathak said, "and it has extremely low latencies." Notable users include The New York Times, Heroku and IMDB.
Amazon RedShift. RedShift is a fully managed petabyte-scale data warehouse service. "With RedShift, you can deploy a data warehouse within minutes," Pathak said. "It's faster, cheaper and easier than other options out there." It is designed for analytic workloads and connects to standard SQL-based clients and business intelligence tools.
Learning more about RedShift was a main draw for at least one conference attendee at this year's AWS Summit. "I came to see some more demos of RedShift," said Yuval Naveh, chief technology officer at Jersey City, N.J.-based Matrix-Exzac, a global consulting and compliance technology systems integration firm.
"We're not using RedShift now, but I think we will be soon," Naveh said. "The classic data warehouse is reaching a point where it just cannot deal with big data; the time window is constant, but volume is increasing. RedShift can help with that, because it's still a data warehouse but it's really fast, compliant with SQL and super-fast for queries."
Follow us on Twitter at @CloudAppsTT.