Should Data Centers Think?

As cloud computing becomes the information technology mainstream, data center technology is accelerating at a breakneck speed. Concepts like software define infrastructure, data center analytics and Nonvolatile Memory Express (NVMe) over Fabrics are changing the very nature of data center management.  According to industry research firm IDC, organizations will spend an additional $142.8 billion oninfrastructure for both public and private cloud environments in the next three years (2016-2018) to boost efficiency and business agility.

To support this rapid evolving space, Intel announced a “Cloud for All” initiative last year in order to help businesses get the most out of their cloud infrastructure. Specific goals for this initiative include:

• Investing in the ecosystem to accelerate enterprise-ready, easy-to-deploy software defined infrastructure (SDI) solutions;
• Optimizing SDI solutions to deliver highly efficient clouds across a range of workloads by taking full advantage of Intel platform capabilities; and
• Aligning the industry and engaging the community through open industry standards, solutions and routes to market to accelerate cloud deployment.

As cloud infrastructure management is moving towards these new management paradigms, those at the leading edge are exploring how to make data center’s think for themselves. Industry leaders like Dr. Brian Womack, Director of Distributed Analytics Solutions in Intel’s Data Center Solutions Group, and Das Kamhout, Senior Principal Engineer at Intel are learning how to use data, artificial intelligence frameworks and machine learning to create data centers that think for themselves. Two key components of their vision are SNAP and TAP.

SNAP is a powerful open data center telemetry framework. It can be used to easily collect, process, and publish telemetry data at scale. It enables better data center scheduling and workload management through access to underlying telemetry data and platform metrics. The framework greatly improves system administrator control of the intelligent use of data center infrastructure in cloud environments by:

• Empowering systems to expose a consistent set of telemetry data;
• Simplifying telemetry ingestion across ubiquitous storage system;
•  Improving the deployment model, packaging and flexibility for collecting telemetry;
• Allowing flexible processing of telemetry data on agent (e.g. machine learning); and
• Providing powerful clustered control of telemetry workflows across small or large clusters.

Trusted Analytics Platform (TAP) makes the SNAP telemetry usable by providing the tools, components and services necessary in the creation of advanced analytics and machine learning solutions. TAP makes these resources accessible in one place for data scientists, application developers and system operators. An open-source software platform optimized for performance and security, TAP simplifies solution development through the delivery of a collaborative and flexible integrated environment.

With TAP, Interactive analysis, modeling and algorithmic process flows on any type of raw data, streaming in real-time or batch data, is possible using either a GUI or a text-based shell. These models and flows can be used for batch processing or be integrated into applications. TAP includes REST APIs usable by any web-capable language (e.g., Python, Java, PHP, Ruby, Javascript) over HTTP, as well as a Python API, for server-local access. It operates on most data stores and file systems, including cluster federations that can enable data sharing (with security). The integrated operations management tools in TAP allow monitoring and control from top to bottom. In support of trust, TAP Security follows layered security and deep defense principles to provide transparent encryption and decryption, as well as fine-grained access authorization, based on a variety of authentication mechanisms and assurance levels.

Used in combination, SNAP and TAP could be used to make sentient data centers a reality.

Visit Chip Chat to hear more more about creating a data center that thinks for itself!

Follow me at https://Twitter.com/Kevin_Jackson