Smorgasbord of announcements from IBM includes cloud for deep learning, cloud for compliance, and cloud for scientists
IBM has made three announcements concerning its cloud offerings, which have been tailored for particular vertical industry segments.
The first concerns a private cloud version of its collaborative development environment, called the Data Science Experience. This is geared towards the scientific community.
The second concerns a deep learning cloud that is claimed to significantly reduce deep learning training time thanks to the inclusion of GPU technology.
And the third announcement centres around a more controlled offering for the enterprise market, where compliance and security challenges can often hinder the adoption of cloud technology.
The launch of a private cloud version of IBM’s collaborative development environment, called Data Science Experience Local, is aimed at helping businesses and scientists working with sensitive data.
“The workplace is designed to help data scientists more easily and quickly collaborate on analytic models that developers can use to build intelligent applications,” said IBM, citing the problem of data scientists having to work with mountains of data that are pulled into servers and data centres.
This means that for some some organizations, moving that data to the cloud isn’t an option due to such constraints as volume, siloed systems and compliance requirements.
IBM says that its Data Science Experience Local, is completely self-contained and resides within an organisation’s own servers and data centre.
It also comes with all the necessary software to run and manage the development environment, including local installations of Apache Spark and Object Storage in addition to Data Science Experience services.
It runs in Kubernetes, an open source cluster manager.
“Industries from healthcare to financial services, demand greater rigor around the ingestion, sharing and analysing of their critical data,” said Rob Thomas, General Manager, IBM Analytics. “With the new local version of the Data Science Experience, data scientists now have a collaborative development environment from within a private cloud setting to quickly and securely extract valuable insights in order to make strategic, data-driven decisions.”
The second cloud announcement the IBM Cloud geared towards advance performance often associated with training deep learning models.
Big Blue says it has the benchmarks to prove the NVIDIA Tesla P100 GPU accelerators on the IBM Cloud (revealed in February) can provide up to 2.8 times more performance than the previous generation NVIDIA Tesla K80 for certain test cases.
It believes these benchmarks means that it can reduce deep learning training time by up to 65 percent compared to NVIDIA Tesla K80 GPU accelerators on IBM Cloud.
This is needed as demand for AI applications is growing rapidly, and IBM said that deep learning techniques are a key driver behind the increased demand for and sophistication of AI applications.
However, training a deep learning model to do a specific task is a compute-heavy process that can be time and cost-intensive.
“Innovation in AI is happening at a breakneck speed thanks to advances in cloud computing,” said John Considine, general manager, cloud infrastructure services at IBM. “As the first major cloud provider to offer the NVIDIA Tesla P100 GPU, IBM Cloud is providing enterprises with accelerated performance so they can quickly and more cost-effectively create sophisticated AI and cognitive experiences for their end-users.”
The third and final cloud announcement from IBM is aimed at the enterprise market, where precise control is often needed over cloud environments for security and compliance reasons.
The new virtual server offering is called Dedicated Hosts, and will supplement IBM’s existing single-tenant offering by giving clients maximum control of their cloud workload placement and post-deployment management.
Dedicated Hosts is actually a physical server with workload capacity entirely dedicated to a single client’s use. This makes it ideal for businesses migrating to a public cloud environment, but which still require exact control and visibility into where their data and workloads reside.
“Enterprises turn to IBM because they want cloud infrastructure that is flexible and easy to scale, while providing higher levels of control and visibility to help as they meet their regulatory and compliance requirements,” said John Considine, general manager, cloud infrastructure services at IBM.
IBM plans to make Dedicated Hosts available in June 2017.