MapReduce begin encountering woes in extrapolating information as big data got bigger, so Google revealed its replacement for the decade-old analytics software and announced Cloud Dataflow at Google I/O 2014.
Cloud Dataflow's build was constructed using Google's Flume, a Java library used to create pipelines for parallel computation, and its MillWheel framework, a platform that facilitates the construction of low-latency applications.
Greg DeMichillie, director of product management for Google's Cloud division, stated in a June 25 release that Cloud Dataflow was designed to pick up where MapReduce's batch processing left off, by implementing the ability to process streamed information as well as input that has been clustered.
"Cloud Dataflow makes it easy for you to get actionable insights from your data while lowering operational costs without the hassles of deploying, maintaining or scaling infrastructure," stated the release. "You can use Cloud Dataflow for use cases like ETL, batch data processing and streaming analytics, and it will automatically optimize, deploy and manage the code and resources required."
Along with news of Dataflow, DeMichillie's release also stated that Google was releasing new diagnostics tools for its cloud platforms. In a recent interview, DeMichillie stated that Dataflow was the spearhead at the tip of Google's thrust to push big data onto its cloud platforms.
"Dataflow is the first of what we're doing to help people make sense of the data that they have," DeMichillie said in an interview. "They need to focus on their business problems, not managing the infrastructure... Dataflow begins that and we'll continue that."
Some of the cloud platform DeMichille revealed in the June 25 release included Google Cloud Monitoring, Cloud Trace and Cloud Debugger. Built from designs awarded the company in its acquisition of Stackdriver and dubbed Google Cloud Monitoring, Google's new monitoring tool delivers metrics and includes a dashboard that enables administrators to track their systems' health.
Cloud Trace will help administrators visualize the latency an application encounters when processing requests and it enables them to compare the data with previous versions of the software. Meanwhile, Cloud Debugger delivers on the promise of its name and enables developers to debug apps without pulling them from the cloud for downtime.
Google also wants to want the push the development of mobile apps into the cloud and encourage developers to use cloud platforms as the backbones for their apps. DeMichillie referenced the I/O 2014 demonstration of Google Cloud Save, touting the service's ability to relieve developers of having to code their apps on the back-end to take advantage of cloud service.
Though a information on a lot of new cloud services was packed into a single post, DeMichillie assured developers that more details would soon follow.
