testbeds:bridges from research to infrastructure Ashkan sobhani

What is testbeds?  Any project experimenting with new capabilities.  Testbeds are complex combinations of technology and peaple.  Some testbeds are aimed at specific Communities,others strive for more ambitious scale.

Good testbeds  The most good testbeds must be consist of the right balance of scale,component technologies,and coupling between the envisioned capabilities of technology and the need of the target communities.

 developing,testing,and refining a technology are the function of a testbed  The two type of create systems: 1.expansion,biger,faster,… of currently availabe systems. 2.Cannot be acomplished by mere extensions of current technology.

 In the second type for building computational grids,testbeds are critical in at least three ways: 1.Scale of integration 2.Building communities 3.Mitigating risk

History of the Network The grandfather of internet

ARPANET  During the early 1960s, researchers in United States began to develop the concept of communications network that would send information in packets.

ARPANET(cont.)  In 1958, the Advanced Research Projects Agency (ARPA) of the Department of Defense (DoD) was created  DARPA launched ARPANET. The network was designed to help government scientists communicate and share information. It was originally developed to allow researchers to log-in and run programs on remote computers, but it quickly became a tool for sharing information through file transfer, electronic mail, and interest group mailing lists.  Batch processing=>timeshared use

ARPANET(cont.) ARPANET December 1969 From ARPANET Completion Report, BBN, 1978

ARPANET(cont.)  Build an interface message processor(IMP) that would be the building block to a packet-switched network.  Packet-switched network,having been proven viable by ARPANET.

POST-ARPANET NETWORK TESTBEDS  Durring the late 1970s and early 1980s,ARPANET testbeds results made their way to infrastructure.  Influenced the computer industry.  MILNET was created to interconnect defense sites using ARPANET technology.  Schlumberger use ARPANET tech. for its worldwide corpoate network

POST-ARPANET NETWORK TESTBEDS  New companies formed commercialized the technology for example: GTETelenet & Tymnet  The ARPANET testbed concepts and technology led to other government and academic network projects such as: CSNET,BITNET,MFENET,ESNET, NSFNET  The second generation of network was created.

NSFnet  Started off connecting 5 Supercomputer Centers between 1985 and JvNSC, SDSC, NCSA, CTC, PSCJvNSC, SDSC, NCSA, CTC, PSC  The origenal goals of NSFNET program were to use ARPANET tech. to provide infrastructure for supercomputer user.  Grew to connect Universities and Regional Nets  Served as the Internet’s backbone grew even further and faster as the Internet grewgrew even further and faster as the Internet grew

NSFnet(cont.)  A three-layer model 1.A bckbone network 1.A bckbone network 2.Mid-layer network 2.Mid-layer network 3.Campus network 3.Campus network  Use internet protcol (IP,TCP)  The selection of IP & interconnecting LANs =>need for routers in the NSFNET backbone =>need for routers in the NSFNET backbone  Use minicomputer based routers called “fuzzballs”

NSFNET Network Structure Backbone Regional Access Providers Local Access Providers

NSFNET Network Structure Backbone Network Access Providers Internet Service Provider vBNS - Very High Speed Backbone Local Service Provider California Washington New York Chicago Sprint MCIAT&T AOLUUNETPSI Local Providers Local Providers

NSFnet(cont.)

NSFnet(cont.)  Gange in protocols and create translator program(GATED)  Generate monitoring protocols as SGMP  Add the congestion control to protocol  It was soon upgraded to T1 speeds (1.5 Mbps).  In 1989, it was apparent that even greater bandwidth would be needed and it was upgraded to T3 (45 Mbps) in  1995 Internet is served by multiple commercial backbonesInternet is served by multiple commercial backbones NSF shuts the NSFnet downNSF shuts the NSFnet down

NSFnet(cont.)

Gigabit Testbed program  DARPA & NSF & CNRI proposed a major program in high-speed neworks. a major program in high-speed neworks. The initiative would attempt to answer two quastions: 1.How woud a gigabit-per-second network be architected? 2.What would its utility be to end users?

Gigabit Testbed program(cont.)  Five testbeds were formed in  That consist of: 1.CASA: focused on ditributed supercompting applications.the CASA network constructed by using HIPPI switches interconnected by HIPPI- over-SONET at OC-12(622 Mb/s).

Gigabit Testbed program(cont.) 2.BLANCA:included virtual enviroment,remote visulization and multimedia digital libraries.using ATM switches running over 622 Mb/s. switches running over 622 Mb/s. 3.VISTANET:supported the development of a radiation treatment planning application.Using an ATM network at oc-12(622 Mb/s).

Gigabit Testbed program(cont.) 4.NECTAR:focus on the application work involved coupling supercomputers running chemical reaction dynamics. 5.AURORA:Focused on network and computer science issues.Using ATM switches & interconnections at OC- 12(622 Mb/s).

Gigabit Testbed program(cont.)  At the same time, the fact that the testbeds combind research at multiple layers( from hardware to network protocol to middleware to applications)

Gigabit Testbed program(cont.)  The important results of these testbeds are: 1.Feasibility of using ATM in a high- performance network. 2.Successful demonstration of OC-3(155 Mb/s),OC-12(622 Mb/s),OC-48 (2.4 Gb/s). Thus technology deployment within the telecommunications industry was accelerated by the gigabit testbeds.

SYSTEM TESTBEDS  Testbeds have been used in the development of hardware,software,and systems.  Two testbed efforts durring the past several years illustrate the advantages of using existing network technology.  These are significantly different from network testbeds.  Focus on improvement the software technology.

SYSTEM TESTBEDS  These testbeds consist of: 1.I-WAY:attempted to exploit the multiple ATM testbeds for support high- performance application in science engineering. 2.ARIES:was aimed at exploiting existing ATM services to support applications important to the oil industry.

I-WAY  I-WAY applications were classified into five general categories: 1.Distributed supercomputing 1.Distributed supercomputing 2.Remote visualization 2.Remote visualization 3.Collaborative enviroment 3.Collaborative enviroment 4.Distributed supercomputing coupled 4.Distributed supercomputing coupled with collaborative enviroment with collaborative enviroment 5.video 5.video

THE LANDSCAPE IN 1998  Several important network testbeds are worth examining today as well,including the vBNS,AAInet,CAIRN,SuperJANET and CANET*2

TESTBEDS FOR FUTURE  Evolution and Revolution  Getting real user involved  Funding and Organization

Evolution and Revolution  The testbeds cited have proven that evolution is one of the most important characteritics of producing revolutonary results in computing and information technolgy.  In addition,testbeds focused on revolutionary technologies of the day eventually evolved into broad-based classes of applications

Getting Real Users involved  A major advantage of testbeds is to get real users involved  This marriage of user and technology is often the key to success.  Success or refinements in any given testbeds will lead to the definition of the next wave of testbeds which can bring new users,technologies,…

Funding and Organization  Sustained funding for a series of testbeds is essential to realize their potential and,even more important,to attract the users,experiments,researches needed to drive the testbed to meet the aggressive goals.

CONCLUSIONS  Larg and small testbeds have been key contributors to today’s core networking,software,and computing infrastructure.  The successful construction of the national-scale grid enviroments will require careful choices as we select the technologies and communities that will form the testbeds for tomorrow

Tank you. Tank you.