Changes between Version 8 and Version 9 of coin


Ignore:
Timestamp:
Oct 12, 2018, 5:50:42 AM (11 months ago)
Author:
marie@…
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • coin

    v8 v9  
    33'''Introduction'''
    44
    5 We are experiencing a convergence between the concepts of networking and computing, triggered not only by the softwarization of networking functions (SDN, NFV) but also by the evolution of the network architecture. The move to edge distributed computing/networking is also encouraging the development of local networking and computing facilities to support low delay and low loss services that are emerging from AR/VR, autonomous vehicles and intelligent/smart cities. Consequently the idea of a “programmable” network is central to the evolution of the Internet.  Programmable data planes are now available with acceptable and ever better performance. Programmable switches and abstraction such as P4 l in Data Center Networks and the rise of virtual network devices in NFV in both DCN and carrier’s network just confirm this evolution as are the programmable network processing units (NPUs) in traditional routers that also programmable in some degree.
     5Integrating computing and networking has been widely investigated and applied at several network layers in the past. Most notably, “Active Networking” research in the 1990s explored approaches for allowing packets and datagrams flowing through a network to modify the behavior of the network itself. This could be done at several layers, e.g., enabling/modifying transport protocol behavior, configuring or programming link layer functionality
     6upon connection establishment etc.
     7
     8Hence are experiencing a convergence between the concepts of networking and computing, triggered not only by the softwarization of networking functions (SDN, NFV) but also by the evolution of the network architecture. The move to edge distributed computing/networking is also encouraging the development of local networking and computing facilities to support low delay and low loss services that are emerging from AR/VR, autonomous vehicles and intelligent/smart cities.
     9
     10Recent research in network data plane programmability has enabled new ways for relaxing the boundaries between strictly network layer and application layer programmability. For example, programming abstractions such as P4 and more powerful programmable DC switch platforms enable the implementation of different support functions for application layers entities, supporting applications such as DNN (Deep Neural Network) training, frontend KV (Key-Value) caching for skewed and dynamic workloads, and high-performance consensus protocols such as
     11Paxos.
     12
     13In addition, there are scalable stream processing frameworks such as Apache Spark or Apache Flink that apply programmed functions on data flowing in a distributed system. These platforms are typically concerned with guaranteeing certain semantics and providing high reliability and performance by orchestrating the set of functions accordingly. Such as distributed processing platforms are overlays(from a network layer perspective) but also have to deal with flow/congestion control etc. on their respective layer.
     14
     15In parallel, there are approaches for connecting so-called network functions (such as NFV -- Network Functions Virtualization) and derived/related computing/networking models such as CDN and Edge Computing -- which are mostly concerned with setting up and maintaining overlays and virtual networks between application logic in virtual machines etc. Consequently the idea of a “programmable” network is central to the evolution of the Internet. 
     16
    617
    718'''Motivation'''
    819
    9 Distributed computing in the network  provides new opportunities to enhance performance and availability as well as to develop new types of networked applications and systems.
     20Network programmability provides new opportunities to enhance performance and availability as well as to develop new types of networked applications and systems. Looking at different instantiations of integrating computing and networking, the following questions arise:
    1021
    11 Examples include:
     221) Are there common principles, abstractions and mechanisms that can be applied across this range of different types of computing/networking approaches?
    1223
    13 ''1) Data plane performance''
    14 The research community is active in finding innovations that use in-network compute and cache capability and expand the reach of programmable control and data planes to improve the performance of distributed systems. Examples include deep neural network (DNN) training for management, control and adaptation, distributed key-value store, local loss coding (network and application layer codes) and distributed system consensus (for example PAXOS and blockchains). The results show that significant (up to 10x) performance improvements when compared with centralized solution that experience delay and losses due to bottleneck situation.
     242) What are best practices and relevant considerations for computing/networking systems, in particular with respect to previous discussions regarding active networking and end-to-end-arguments?
    1525
     263) Many of the computing/networking systems developed for and by the networking community have been designed from a networking perspective (assuming running applications that need to be connected by
     27connections, overlays, service chains). Is there potential to include design patterns from the distributed computing and applications community (e.g., stream processing) that would allow for a more holistic design, i.e., considering both networking and computing for optimizing the layout of processing functions and distribution of data?
    1628
    17 ''2) Decentralized, lightweight and dynamic computing''
    18 In order to complement the more centralized data centers, there is a wide interest into decentralized computing such as edge/fog/pervasive (ubiquitous) computing that profit from proximity to the end user to support local functionality or enhance DC services.  In parallel, “serverless” and FaaS (function as a Service) are changing the nature of the client-server model. The logic that was at the server side is moving to the client, while “application servers” are decomposed into “functions” that are triggered by client requests. Unlike deploying a server or virtual machine or even a container, the computing load for these “functions” can be lightweight and can instantiated in 100ms level when they are requested.
    19 
    20 All these new trends together are enabling innovative research and even a re-definition of what it means to be “in the network” and providing the tools to develop new networking and computing paradigms in a distributed architecture.
    2129
    2230'''Research Challenges'''
    2331
    2432The research questions that the COIN group wants to address include but are not limited to:
     33
    2534(1) Even within the traditional and "end-to-end argument", will distributed computing in the network provide enough motivation and benefits to justify the introduction of non-forwarding functions into the network?
    2635
     
    4352
    4453'''COIN Objectives'''
     54
     55The main objectives of the COIN RG include:
    4556
    46571)      Understanding the use cases and different types of network programmability and their different characteristics (for example, DC switch programmability vs. distributed/edge computing).