With Google Cloud Next around the corner, Google laid back on their updates the last month. But as always, our Google evangelist is here to bring you all the latest updates on Data!
Data plays the leading role at Google
Again a series of improvements to CloudSQL, Datasets and Pub/Sub. We also highlight the strong connection between SAP and BigQuery. In machine learning, we see the first extensions in Vertex AI and new functionalities in Dialogflow CX.
Pub/sub simplifies the reuse of posts
Since August, with Topic Retention you have a new simple and more powerful way to save and replay posts published in Pub/Sub. Previously you had to save messages that you configure and pay separately in each subscription. Now when you enable topic retention, all messages sent to the topic within the chosen retention period will be accessible to all the topic’s subscriptions, without incurring storage costs when you add subscriptions. In addition, posts are kept and can be replayed even if there are no subscriptions associated with the topic at the time the posts are published.
In addition, the new feature extends the existing Pub/Sub search functionality: message replay is no longer limited to the confirmed messages of the subscription.
Cloud SQL PostgreSQL Performance Gain with Linux Huge Pages
As of now, Cloud SQL voor PostgreSQL supports Linux Huge Pages by default. This means that if your instance has more than 2 GB of RAM, your PostgreSQL instance will use Linux Huge Pages for shared memory. Or concretely that your shared memory in your Cloud SQL for PostgreSQL instance now uses 2 MB pages instead of the normal 4 KB pages. This leads to improved memory usage, freeing up more memory for database operations, and improved performance through more efficient use of the Translation Lookaside Buffer (TLB).
A growing number of datasets available for analysis and ML
With Google Cloud Datasets, as a user in Google Cloud Platform Marketplace, you get access to a growing inventory of specific data collections that can help you support and strengthen your analytics and ML models, as well as regularly updated best practices that tell you how to get the most out of these datasets. To keep users informed about new available datasets, you can refer to the blog below.
SAP – BigQuery combination, a strong story
Today no one disputes the importance of data and especially the insights from data as the basis for conducting a policy. However, it remains a challenge to distil those insights from all the data that you produce as a company. Finding the right tooling is certainly not one of the least tasks.
In a recent study by IDC, they looked at the combination of SAP (as an important but not unique source of the data) and BigQuery (as a platform for the analysis). The study looked at large companies (average annual revenue of $1.8 billion, 36,000 employees and 1.5 PB of data).
The study found that the combination produced visible gains due to three elements:
- Increased productivity (such as accelerating production reports by 67%);
- Improved IT staff productivity (e.g. by eliminating a series of administrative tasks);
- Lower infrastructure costs (e.g. by grouping data on one platform);
See the report: The business value of BigQuery for SAP
Google makes the difference with AI/ML offerings
Deliver AI predictions to your customer safely and quickly
One of the biggest challenges in using machine learning models is delivering predictions in real time. Just think of an online store that generates recommendations for its customers for a specific product or a food service company that has to estimate its delivery time. In all these cases it is crucial to be able to deliver results with low latency. With the Private Endpoints in Vertex Predictions launched on Vertex AI in August, you can set up a private connection via VPC peering to talk to your endpoint without your data ever going out on the public internet, resulting in increased security but also lower latency for your online predictions.
Dialogflow CX, Google’s chatbot gets new features
With six new features in preview, Google aims to further strengthen the position of Dialogflow CX as a virtual chatbot agent for businesses.
With these features, customers can improve their end-user conversation experience and achieve greater security and deployment. In addition, there are some console improvements and built-in support to make bot building more efficient and thus take the user experience to the next level.
The new features relate to the following domains:
- Streaming a partial response;
- Private network access to Webhook targets;
- Search options in the console;
- System Function Support:
- Ongoing testing and implementation;
- Changes in the history logs;
Read more: Six new features in Dialogflow CX
See you next month!