ra2 studio - Fotolia
Google Cloud Platform has fired another shot at rivals AWS and Microsoft Azure with the general availability of Google Cloud Firestore -- a fully managed NoSQL document database, with particular relevance for mobile and IoT-centric applications.
Now generally available after 15 months in beta, Google Cloud Firestore is a successor of sorts to Realtime Database, which is part of Google's Firebase mobile and web application development platform. Developers can use their normal tool chain when they interact with Cloud Firestore, which also includes an iOS software development kit and web SDK.
Other supported languages include Node.js, Ruby, Python and Java. Cloud Firestore also offers offline capabilities through the Android SDK, so apps continue to work in areas with unreliable network coverage.
Google Cloud Platform still supports Realtime Database for now, but Cloud Firestore improves upon Realtime with richer query capabilities and better scalability, according to a GCP FAQ.
Google Cloud Firestore was available in three regions during the trial, but it has since expanded to nine more regions and added a multi-regional instance option, GCP said. Google will offer new regional instance pricing of up to 50% off, as well as service-level agreements of five nines for multi-region deployments and four nines for single-region ones.
Customers face choices, as cloud NoSQL options proliferate
There are close similarities between Google Cloud Firestore and AWS DocumentDB and Azure Cosmos DB, and customers likely will gravitate toward wherever they already have cloud workloads, said Nick Heudecker, a Gartner analyst.
GCP has a strong story to tell, with multi-region resiliency and the robustness of its underlying network architecture, Heudecker said.
On the general topic of NoSQL versus relational data stores, there are important points to consider.
"Document stores are the top reason I draw the distinction between 'structured data' and 'structured databases,'" said Curt Monash, an analyst at Monash Research in Acton, Mass. "Relational databases have a lot of structure so that the data doesn't have to. Document databases are the other way around."
Determining which one is better depends largely on two related factors, Monash said. First, consider whether you want to combine the jobs of programmer and database administrator or separate them. "NoSQL prospered where the separation wasn't wanted," he said.
If a company wants to run many applications against the same data set, relational databases offer advantages, he said.
Database management systems with support for occasionally connected scenarios -- such as the ones offered by Google Cloud Firestore and Sybase's SQL Anywhere database -- are valuable, but customers have come to expect always-on connectivity, Monash said
"I hope that changes, because forced always-on connectivity makes it easier for vendors to enforce practices that are bad for users in terms both of privacy and of flexibility as to what software they use," he said.
Hawkin Dynamics finds growth engine in Cloud Firestore
Hawkin Dynamics, which makes force plates and related mobile software to measure athletic performance, is an early customer of Cloud Firestore.
Cloud Firestore and its close proximity to other GCP products solved a lot of problems for Hawkin Dynamics, said Chris Wales, CTO of the startup in Westbrook, Maine.
"You can just sort of reach for the next thing that you need as you grow," he said. Examples of this include GCP's authentication SDK and Cloud Functions, he said.
Hawkin's hardware and software objectively quantifies the amount of force produced by athletes in motion. It's used by strength and conditioning coaches, mostly at the professional and elite college levels, to track the progress of athletes.
Force plates are not new, but Hawkin's goal is to build a system that tackles thorny problems faced by previous iterations of the idea. For example, some force plate systems use wired connections to tie in athletes, and this isn't ideal, he said. Hawkin's mobile app circumvents the issue.
Hawkin initially began to build its mobile app with Realtime Database, but switched to Cloud Firestore because more advanced features would be easier to implement, Wales said. Specifically, Hawkin wanted to move from an app that could show results of force plate tests to one that could support more ambitious goals, such as longitudinal analysis of an athlete's performance over time.
Meanwhile, GCP's auto-scaling capability solves other problems, Wales said. "It's like having your own ops team out there doing your work on your behalf."
GCP's multi-region and failover features also mean Hawkin can add customers around the world without concern for how its products will perform, Wales added.
Wales also highlighted Cloud Firestore's offline capabilities. "In a gym, you don't want coaches worrying about whether the Wi-Fi is working or not," he said.