Open Cloud Test Bed Bolsters Big Data Innovation
On Friday, in the Western Massachusetts town of Holyoke, Governor Deval Patrick and officials from industry, government and academia joined together for the official launch of a $3 million capital investment, known as the Massachusetts Open Cloud (MOC) project. The public-private initiative will establish a new cloud computing infrastructure to serve as a virtual laboratory for big data researchers and innovators across the state. One of the first aims of the MOC Project will be to study the feasibility of operating HPC applications in this open cloud environment.
Although the MOC will begin as a research collaboration and testbed, the ultimate goal is for it to become an independent non-profit entity. Guiding its direction as it evolves from prototype to production to self-sustaining operation will be a wide range of stakeholders – a mix of academic, industry, government and non-profit members. The initial $3 million in funding is being provided by the Commonwealth of Massachusetts under a Mass Tech Collaborative Matching Grant Award with another $16 million in matching funds coming from a mix of federal, industry and philanthropic sources.
The physical resources for the cloud, which will be implemented using an OpenStack framework, will be hosted at the Massachusetts Green High Performance Computing Center (MGHPCC), the Holyoke, Mass., datacenter that was created in 2012 by Boston University (BU), Harvard University, MIT, Northeastern University, and the University of Massachusetts (Umass). The 90,000 square foot, 10 megawatt facility is located on an 8.6 acre former industrial site just a few blocks from City Hall in Holyoke, Mass.
The five founding universities are also principal MOC partners along with Massachusetts Green High-Performance Computing Center (MGHPCC) and Oak Ridge National Laboratory (ORNL). Boston University’s Hariri Institute for Computing and Computational Science and Engineering is leading the project with operational leadership coming from Harvard University, development leadership from Northeastern University, community building from MIT, and related research falling under the purview of all five universities.
The initiative is based on the “Open Cloud eXchange” model, where hardware, software and other services can be supplied, purchased and resold by many participants, ranging from existing providers to startup innovators. As with popular public clouds from vendors like Amazon and Google, MOC provides access to massive off-site computational resources, but unlike those proprietary clouds, where all of the technology is controlled by a single provider, MOC uses an open and customizable approach to the design and operation of cloud computing.
It makes sense that a program designed to spur innovation would itself be on the cutting-edge. “The MOC will be the first realization of this model,” reports Orran Krieger, director of the Cloud Computing Initiative at the Hariri Institute. “If it’s successful, we expect other clouds to follow our model, fundamentally changing the nature of cloud computing.” Krieger is also a College of Arts & Sciences research professor of computer science at Boston University.
Krieger and fellow Boston University researcher Azer Bestavros were instrumental in designing the model, and seeing it to fruition.
“The Open Cloud Exchange (OCX) is envisioned as a public cloud marketplace in which many stakeholders, rather than just a single provider, participate in implementing and operating the cloud. This ecosystem would enable innovation from the broader academic and industry communities, resulting in a much healthier and more efficient cloud marketplace,” the duo write in a recent paper.
They make the case that a closed cloud stifles innovation, whereas an open cloud promotes it. With a comparison to open source software, the authors are confident an open cloud will also enhance security.
“Th[e] single-provider model results in strong homogeneity in cloud providers’ offerings,” they write. “This not only limits research but also results in a monoculture that’s susceptible to security threats. Security concerns also arise because public clouds are designed under the assumption that the cloud operator is fundamentally trusted. No documented technologies or policies keep a cloud provider, or even a disgruntled employee, from instrospecting on a customer’s network traffic, computers or even private datasets. We argue that an open cloud would deliver some of the same long-recognized security benefits as has open source software development.”
According to the project’s website, the MOC has two overarching goals:
- To create an improved computing resource for cloud and big data users in the Commonwealth.
- To create a new model of cloud computing that enables research and technology companies to innovate and profit in the cloud and big data sectors.
Major tasks and milestones have been divided into three main categories:
- Deploying, operating, and maintaining a production cloud service with technology partners.
- Enhancing OpenStack to enable multiple competing providers to participate in a shared cloud.
- Working with a broad industry and academic community to enable new workloads and users to exploit the cloud.
The MOC project is already enjoying strong industry support. Founding vendor partners include Cisco, DataDirect Networks, EMC, SGI, Red Hat, Juniper Networks, Canonical, Dell, Intel, Mellanox, Brocade, Mathworks, Plexxi, Cambridge Computer Services, Enterprise DB, and Riverbed.
DDN’s role in the project will be to contribute its Web Object Scaler (WOS) software to enable a low latency object storage service. DDN’s Chief Technology Officer Jean-Luc Chatelain notes that while the initial interface to the WOS cloud will be through OpenStack APIs, the company expects to be able to test multiple interfaces and configurations at large scale. (WOS can host up 32 trillion objects, according to the company.) In an interview with HPCwire, Dave Fellinger, DDN Chief Scientist, emphasized the importance of data management in a federated environment to enable collaboration. The MOC Project has a long-term goal of federating with other datacenters, and DDN’s WOS platform has been designed to enable this service, without involving servers or any external management devices.
As Fellinger further explains, WOS was designed to be an efficient means of data replication and data recovery. The software establishes on each storage node a demon that runs across the entire storage cluster, tracking and recovering data on a peer to peer basis. If a disk drive fails, the node will retrieve the objects and replace them automatically without any server involvement.
SGI is another company that’s excited to be collaborating on the Massachusetts Open Cloud project. “SGI’s computational infrastructure is purpose-built to handle workflows that lie at the intersection of HPC and Big Data – High Performance Data Analysis is a growing area for data scientists in both academic and commercial settings, and we are proud to help Massachusetts grow its leadership efforts in these areas,” Jorge Titinger, president and CEO of SGI states. “As a result of this collaboration SGI expects to contribute to further advances in Knowledge Discovery, the science of extracting exceptional insights from very large and continuously dynamic data sets, paving the way for real Big Data innovation.”
Governor Patrick also announced the release of the 2014 Mass Big Data Report, which identifies and outlines the opportunities for innovation and growth within the big data industry, providing strategic recommendations for policymakers. According to findings in the report, the global big data market is expected to hit $48 billion by 2017, up from $11.6 billion in 2012. Hardware and services are expected to continue to account for the greatest share of revenue, however, the fastest growing sector, according to the report, is likely to be in big data-enabled applications. In Mass., big data applications are especially well-represented in healthcare, life sciences and financial services, and local firms are seeking to fill as many as 3,000 jobs related to these fields over the next 12 months.
“Investment in the Massachusetts Open Cloud will help keep our Commonwealth at the forefront of big data research nationally, expanding opportunities for innovators to build advanced cloud computing solutions,” said Pamela Goldberg, CEO of the Massachusetts Technology Collaborative. “As cited in the 2014 Mass Big Data Report, we must continue developing cross-sector collaborations like the Massachusetts Open Cloud, in order to spur innovation and foster industry growth.”