Frequently Asked Questions

The Massachusetts Open Cloud (MOC) is a new public cloud project announced by Governor Patrick April 25th.  It is a collaboration between the state, local universities and industry. The MOC will stand apart from existing single-provider clouds by enabling multiple entities to provide (rather than just consume) computing resources and services on a level playing field. Companies researchers and innovators can expose hardware or software services through the MOC, make them available to a large community of users, and derive revenue from doing so.  In doing so, the MOC will provide unprecedented visibility and control to users of the cloud, allowing multiple users to simultaneously develop services that meet their specific needs while being isolated from both a performance and security perspective.

The MOC is based on the concept of an Open Cloud eXchange (OCX). In an OCX, multiple vendors, multiple solution providers, and multiple resellers and aggregators of services can all set up shop in the MOC. Because of this diversity, unlike a closed cloud, customers would not be locked in — they can choose, compare, and switch between different offerings. The MOC, as an open cloud, stands in sharp contrast to the monopoly of closed clouds where choices are limited and competition is almost non-existent.

The MOC is conceived as a marketplace where different “parties” can contribute assets in such a way that they can be combined in creative/innovative ways to offer solutions or services that are hard (and often impossible) to develop in a “closed” cloud. These assets can be in the form of hardware, software, or data services, and they can range from individual offerings to aggregates. In the MOC ecosystem, a hardware vendor may offer a specific storage solution; a security company may add value to this solution by reselling it with additional security features; and an integrator may stitch a number of such solutions into offerings with various cost/performance/security thresholds. In that same ecosystem, owners of various data sets may allow innovators to develop platforms that monetize access to their data, thus allowing other innovators to leverage the diversity of data sources for novel applications. The MOC is in essence the convener of all these players; it is the “platform” that catalyzes innovation and hence economic development.

In the MOC setting, the owner of a data asset (say the city of Boston) could store its data on storage devices provided by a hardware vendor (say EMC), allowing an entrepreneur (say a student), with a novel idea that leverages this data, to develop an app that can be sold in the App Store. Of course, the entrepreneur will have to pay Dell for using their computers, and will have to pay the city of Boston for the right to access their data; the city of Boston will have to pay EMC for using their storage; and both EMC and Dell will have to pay the data center (say MGHPCC) for the space and power that their devices use. This marketplace where multiple stakeholders can contribute different assets and where these assets are bought and sold is what is unique about the MOC and what stands in sharp contrast to today’s closed clouds stood up by a single provider.

The MOC is envisioned as evolving to a self-sustaining status, and hence be economically viable, in three years. The $3M funding from the Commonwealth will help our BU-incubated project team bootstrap the development of the software infrastructure that will get the MOC to that point of being economically viable. Also, the $3M funding from the state allows us to leverage significant contributions from companies totaling another $16M of mostly in-kind support e.g. donations of hardware and operational staff.

By virtue of being “open,” any institution/organization/individual can access/leverage the MOC. Realistically, however, we anticipate that until the MOC is fully developed (which we expect will take 2-3 years), most of our “customers” will be those using the cloud for research purposes as opposed to for commercial applications. Our plan is for the MOC to cater to any and all types of applications and uses.

The MOC is based on the concept of an Open Cloud eXchange (OCX).  In an OCX, the “ultimate paying customer” is unlikely be the direct user of OCX services exposed through the various layers of the cloud stack. Rather, the user of OCX services is likely to be an “intermediary” who stitches together various services into a platform or solution to be offered to the ultimate paying customer. Indeed, a major advantage of the OCX model is that it makes the cloud a friendly environment for intermediaries (integrators, brokers, aggregators, resellers, etc.) which is not the case in traditional single-provider public clouds.

Current cloud offerings expose a constrained set of resources, e.g., a few differently sized virtual machines via AWS. The MOC”s OCX-based architecture exposes resources from the raw hardware up through application-layer services, providing a richer environment for developers and solution providers. For example, the MOC will provide and expose the capabilities of more diverse hardware, e.g., GPU clusters, HPC machines, gateways to other facilities. Also, it will enable significantly greater control of its services than what is typically possible with commercial providers. For example, the MOC will allow the deployment of experimental or special-purpose hypervisors; it will allow the development and use of different aggregation and pricing mechanisms; and it will allow the development of new services, such as checkpointing and highly-elastic provisioning of services.

500 TBytes on 10 servers – Lenovo

18 high end servers with 4 TBytes of SSD – Intel

48 servers, 2 TOR Switches – Cisco

18 high end TOR switches – Brocade