Read-replica Support for Apache HBase

You can now create read-replica Apache HBase clusters pointed to the same underlying HBase tables in Amazon S3 on Amazon EMR release 5.7.0. Apache HBase is a distributed, non-relational database built for random, strictly consistent realtime access for tables with billions of rows and millions of columns. By using read-replicas, you can increase availability by creating HBase clusters in different Amazon EC2 Availability Zones that read from the same dataset in Amazon S3.

Continue reading

Amazon EMR now supports launching clusters with custom Amazon Linux AMIs

You can now create Amazon EMR clusters with custom Amazon Machine Images (AMI) running Amazon Linux. This enables you to preload additional software on your AMI and use AMIs that you customize and control. You can also encrypt the Amazon EBS root volume of your AMIs with AWS Key Management Service (KMS) keys. Additionally, you can now adjust the Amazon EBS root volume size for instances in your Amazon EMR cluster. Previously, this was fixed at 10 GiB.

Continue reading

Coming Soon: Microsoft Azure at Illinois

The University of Illinois at Urbana-Champaign has signed a contract with Microsoft to be able to provide access to Microsoft Azure cloud resources. Ultimately, the goal is to create a system similar to the current Amazon Web Services at Illinois offering where University staff and researchers can procure Azure resources while staying aligned with the University’s needs. Now that we have a contract in place, Technology Services is currently working through the logistics required to provision Microsoft Azure resources and re-bill through the University. Continue reading

S3 Presigned URLs

Amazon S3 has been in the news lately:

Top Defense Contractor Left Sensitive Pentagon Files on Amazon Server With No Password

The RNC Files: Inside the Largest US Voter Data Leak

Cloud Leak: How A Verizon Partner Exposed Millions of Customer Accounts

S3’s default configuration does not allow public access to the contents of a bucket, but these stories all feature bucket or object permissions that were open to the world. It’s evident that it’s a common mistake, but how can we avoid it? Continue reading