• SAHARA Project
  • Oasis Sub-Project
• RAD Lab Project
• ICEBERG Project
• Endeavor Project

SAHARA Project

SAHARA Project Goal

The goal of the SAHARA Project at UC Berkeley is to understand how to create end-to-end telecommunications services with desirable and predictable properties, such as performance and reliability, when provisioned from multiple and independent service providers. SAHARA is developing a new architecture for future telecommunications services that supports the dynamic confederation of sometimes collaborating and sometimes competing service providers. SAHARA's first effort in this direction, the Clearinghouse Architecture, provides a resource management system based on predictive resource reservations, traffic-matrix admission control, and group policing for detecting malicious flows. SAHARA is extending this work in the direction of more general application of economic mechanisms, such as dynamic auctions, for resource allocation problems in multi-provider telecommunications service architectures.

Motivation: The Existing Operator Model is Failing

The expense of deploying Third Generation (3G) Telecommunications Systems will be huge. The European auctions for 3G spectra are likely to exceed $150 billion, with $45 billion already committed in Germany and $35 billion in the United Kingdom (U.K.). Equipment outlays are likely to match these spectrum expenses. And this is all before first revenue, without a clear understanding of the kinds of new services and applications enabled by 3G bandwidths for which subscribers will pay. Cheap (core) network bandwidth and the highly competitive environment brought about by widespread liberalization of the telecommunications sector is simultaneously driving bandwidth prices towards zero while yielding a financial crisis for the operators!

There is a growing recognition that highly integrated “all things to all people” telecommunications companies, like AT&T or British Telecomm (BT), on the one hand provide no effective economies of scale and on the other have encumbered very large debt in pursing their integrated visions.

Helpful SAHARA Documents

1. SAHARA First Winter Retreat (January 2002) by Randy Katz (ppt)
2. SAHARA Second Winter Retreat (January 2003) by Randy Katz (ppt)
3. SAHARA Design Review by Randy Katz (ppt)
4. The SAHARA Project: Composition and Cooperation in the New Internet by Randy Katz (ppt)

The Oasis Sub-Project

The OASIS sub-project presents a draft architecture for implementing application-specific "in the network" functionality within the coming generation of programmable network elements (PNEs).

PNEs are network routers that perform complex, diverse, and configurable operations on packets in the routing fast path with minimum latency. PNEs are characterized by being able to classify, infer, and act upon packets using stateful information from both the control and data planes. PNEs represent a paradigm shift from today’s fixed-function network routers to a more dynamic and application-aware packet processor that can interact and cooperate with other PNEs. The promise of PNEs is not only to accelerate and enhance existing network-intensive applications, but and also to enable new types of in-the-network functionality with increased reliability and reduced cost.

The goal of OASIS is to provide a description of architecture in sufficient detail to make it suitable for external review and feedback.

Helpful Oasis Documents

1. SAHARA: An Active Networking Testbed for Storage by Mel Tsai (ppt)
2. From SAHARA to OASIS (ppt)
3. SAHARA - The OASIS Group (paper)
4. OASIS: Overlays and Active Services for Internetworked Storage by Randy Katz (ppt)

RAD Lab Project

The vision of UC Berkeley's RAD Lab is to enable one person to invent and run the next revolutionary IT service, operationally expressing a new business idea as a multi-million-user service over the course of a long weekend. By doing so, RAD Lab hopes to enable an Internet "Fortune 1 million".

To do this, RAD Lab will systematize what has become the de facto standard process for developing, assessing, deploying, and operating such services, by bringing to bear powerful techniques from statistical machine learning (SML) as well as recent insights from networking and distributed systems.

The RAD Lab platform is the modern datacenter. RAD Lab sees the “datacenter operating system” as a split between virtual machines to provide the OS mechanism and SML to provide the overarching policy. To inform the SML policy maker, RAD Lab provides tools that collect sensor data from all the hardware and software components of the data center. To provide actions for the policy maker to take, RAD Lab provides actuators to shutdown, reboot, or migrate services inside the datacenter. Additional technologies to fulfill the vision include workload generators and application simulators that can record behaviors of proprietary systems and then recreate them in a research environment.

Helpful Documents

1. Berkeley RAD Lab: Technical Vision (RADS Retreat, June 2005) (ppt)
2. Berkeley RAD Lab: Research in Internet-scale Computing Systems (ppt)
3. Berkeley RAD Lab: Robust, Adaptive, Distributed Systems (ppt)

Telecommunications networks are migrating towards Internet technology, with voice over IP maturing rapidly. The ICEBERG Project at UC Berkeley believes that the key open challenge for the converged network of the near future is its support for diverse access technologies (such as the Public Switched Telephone Network, digital cellular networks, pager networks, and IP-based networks) and innovative applications seamlessly integrating data and voice. ICEBERG is seeking to meet this challenge with an open and composable service architecture founded on Internet-based standards for flow routing and agent deployment. This enables simple redirection of flows combined with pipelined transformations. These building blocks make possible new applications, like the Universal Inbox. Such an application intercepts flows in a range of formats, originating in different access networks (e.g., voice, fax, e-mail), and delivers them appropriately formatted for a particular end terminal (e.g., handset, fax machine, computer) based on the callee's preferences.

The design of the ICEBERG architecture is driven by the following types of services:

• Any-to-any communication services:
  Any-to-any communication refers to the ability to support communication between all types of devices effectively. For this purpose, ICEBERG employs the Automatic Path Creation service (APC) which composes the data path between the endpoints of any type through mix-matching operators (units of computation) and connectors (transportation abstraction between operators). transport between operators. The use of the APC simplifies extending ICEBERG to new devices, which is merely a matter of introducing new codecs or data transformation operators. By encapsulating the data flow/path management within the APC service, it reduces the complexity of the signaling protocol (the protocol for establishing and maintaining communication sessions).

To enable any-to-any communications, integrated communication systems also need a component that inter-works with various networks for signaling translation and packetization. In the project's control architecture, this component is called an ICEBERG Access Point (IAP).

The MediaManager Service provides multi-modal access to the user's different mail repositories. It can do intelligent transformations such as summarizing a voice-mail. It uses a Transcoder Service for the data transformation.

• Personal mobility services:
  Personal mobility means treating people, rather than devices, as communication endpoints. Every person using ICEBERG has an ICEBERG unique ID. The mapping between the specific device ID and unique ID is done through the Name Mapping Service (NMS). For this release, ICEBERG uses the Open LDAP directory service.

• Communication service customization:
  To allow end users to customize their communication service (such as when they want to be called, on what device, under what condition, and by whom), ICEBERG uses a Preference Registry (PR) to store and manage user preferences. ICEBERG's philosophy here is that callee has the total control on how they can be reached.

User activity-driven services: ICEBERG also aim to support a new kind of communication service based on user activity. This type of service generalizes the location-based services that have appeared in many other systems. Instead of customizing the communication service based on the current user location, ICEBERG allows the current user behavior (such as "I am talking to an important person") to be tracked and used for customization. Users control what behaviors are tracked as a way to control privacy. The Personal Activity Coordinator (PAC) performs the tracking.

Helpful Presentations

1. ISRG Review (Opening Talk at January 2000 Retreat) by Randy Katz (ppt)
2. Touring ICEBERG (an Overview and Tutorial) by Helen Wang (ppt)

Endeavor Project

The Endeavor Project mission is to achieve nothing less than radically enhancing human understanding through the use of information technology, by making it dramatically more convenient for people to interact with information, devices, and other people. Endeavor will achieve this by developing a revolutionary Information Utility, able to operate at planetary scale. To validate the architecture, Endeavor will stress it under demanding applications for rapid decision making and learning. In addition, Endeavor will develop new methodologies for the construction and administration of systems of this unprecedented scale and complexity. Endeavor's success will be measured by how effectively our architecture actually amplifies and leverages human intellect.

The focus of the Endeavor Expedition is the specification, design, and prototype implementation of a planet-scale, self-organizing, and adaptive Information Utility. Its key innovative technological capability is its pervasive support for fluid software. That is, the ability of processing, storage, and data management functionality to arbitrarily and automatically distribute itself among Information Devices and along paths through scalable computing platforms integrated with the network infrastructure, compose itself from pre-existing hardware and software components, satisfy its needs for services while advertising the services it can provide to others, while negotiating interfaces with service providers while adapting its own interfaces to meet "contractual" interfaces with components it services. The fluid paradigm will enable not only mobile code, but also nomadic data, able to duplicate itself and flow through the system where it is needed for reasons of performance or availability. The Information Utility will be designed to support, and to integrate with infrastructure services of processing, storage, and information management, a great diversity of Information Devices. These will include radical devices like MEMS-sensors/actuators and other capture and display devices that go well beyond the straight-forward extrapolations of today's server, desktop and portable computers. The Information Utility architecture will be stressed by using it to enable demanding applications that support collaboration and learning in virtual and physically-enhanced activity spaces.