SA1 Network Infrastructure and QoS

1
SA1Network Infrastructure and QoS

• Davide Adami, Franco Davoli - CNIT
• d.adami_at_iet.unipi.it, franco_at_dist.unige.it,
  franco.davoli_at_cnit.it
• Anastasios Zafeiropoulos GRNET
• tzafeir_at_admin.grnet.gr
• DORII-Meeting October 20-22, 2008, Poznan, Poland

2
After DSA1.1

• CNIT has initiated formal contacts with GARR in
  Italy, in order to approach NRENs and create a
  Task Force dedicated to the project needs.
• GÉANT will be formally contacted by the NRENs.
  However, we are keeping informal contacts with
  Richard Hughes-Jones in the UK, from whom we are
  expecting feedback on the basis of DSA1.1.

3
After DSA1.1 (contd)

• GRNET has conducted a study and prepared a
  document regarding the possible monitoring tools
  that could be deployed within DORII and has
  proposed the possible uses of them.
• In the next months GRNET can install some of them
  (PerfSONAR, MRTG diagrams, Weathermap) for
  monitoring purposes and can contribute to the
  deployment of specific services (e.g. VPN) among
  DORII partners. Wherever possible, the same tools
  will be installed by CNIT, as well.

4
After DSA1.1 (contd)

• Both GRNET and CNIT participate in the committee
  that is going to be created within involved NRENs
  and GEANT2, for the examination and solution of
  networking issues related to DORII.
• Our formal request has been addressed to the
  following contact points (indicated by GARR and
  GRNET)
• GARR Marco Marletta ltmarco.marletta_at_garr.itgt
• PSNC - Arthur Binczewski ltartur_at_man.poznan.plgt
• GRNET - Tryfon Chiotis lttchiotis_at_admin.grnet.grgt
• RED.es - Alberto Perez Gomez ltalberto.perez_at_red.es
  gt
• DFN - Martin Wilhelm ltwilhelm_at_dfn.degt

5
SA1 Objectives

• To investigate and define features of the network
  infrastructure, necessary to support a new form
  of cooperative and distributed utilization of
  resources, such as computing facilities and
  unique advanced instruments
• To provide specific networking infrastructure
  delivering advanced services to the scientific
  communities, including QoS and IPv6 support from
  GEANT2
• To select the best services in accordance to
  collected user requirements
• To design procedures able to detect and configure
  the network devices interfaces to better exploit
  the capabilities of the available networks

6
SA1 Milestones, Schedule, PMs
7
DSA1.1

• Analysis of applications and network requirements
  for remote instrumentation infrastructure
• Main Goals of DSA1.1 are
• to provide a classification of remote
  instrumentation services based on network
  requirements
• to identify the QoS requirements of the
  applications that are considered in the project.

8
Roadmap

• 25 October ToC of DSA1.2 Preliminary design of
  the distributed infrastructure over the GEANT
  network
• 15 November Preliminary Version of DSA1.2
• 21 November Final Version of DSA1.2
• DSA1.2 is due at the end of November 2008

9
SA1 Topics

• QoS requirements of the DORII applications
• End-to-end QoS provisioning and users access to
  network features
• QoS support at NRENs level
• Mobility and security support
• GEANT network services selection
• Mapping QoS requirements at application level
  into network services
• Mechanisms to interface users applications with
  the core infrastructure
• Test bed deployment
• Developing Service Level Specifications and
  Agreements
• Monitoring of the network with regard to SLA
  compliance

10
DORII Applications

• Applications that process pre-collected data
• NCSS (sensor network actuators/sensors)
• FLOAT (sensor network CTD sensors)
• GLIDER (sensor network GLIDER)
• OPATM-BFM (MERSEA remore sensing system)
• HORUS (sensor network digital camera, pressure
  and temperature sensors)
• Applications working on data acquired in
  real-time
• EEWS (seismic sensor network)
• SMIWR (sensor network CTD optical sensors)
• ODAES (sensor network SYRMEP, XRD, SAXS)

11
EUCENTRE - NCSS
12
EUCENTRE - NCSS
13
OGS - FLOAT
Platforms (FLOAT Sensors)
1
Satellite
2
Ground Station
7
3
6
5
Storage Element
Instrument Element
Computing Element
4
8
VCR
9
Web Server
User
14
OGS - FLOAT
15
OGS - GLIDER
Platform (GLIDER)
1
Satellite
2
Ground Station
7
3
6
5
Storage Element
Instrument Element
Computing Element
4
8
VCR
9
Web Server
User
16
OGS - GLIDER
17
OGS - OPATM-BFM
18
OGS - OPATM-BFM
19
UC - HORUS
Case 2 Numerical Models with information comin
g from the camera stations
Case 1 Algorithms testing
Instrument Element (Input)
Storage Element
Storage Element Input
FTP server
input
Computing Element
User interface (VCR/gEclipse)
Storage Element Output
20
UC - HORUS
21
EUCENTRE - EEWS
22
Ecohydros - SMIWR
23
Ecohydros - SMIWR
24
ELETTRA - ODAES
SAXS (Small Angle X-ray Scattering)
25
ELETTRA - ODAES
SYRMEP (SYnchrotron Radiation for MEdical Physics
)
26
ELETTRA - ODAES
XRD (X-Ray Diffraction)
27
ELETTRA - SAXS, SYRMEP, XRD
28
DORII Applications

• Applications that process pre-collected data
• Communication Paradigm Point-to-Point
• Communication Mode
• Sensor Network Unidirectional
• VCR, CE, SE, IE, (Web Server) Bidirectional/Symme
  tric
• QoS Requirements
• Bandwidth 1 Mb/s (NCSS) 100 Mb/s (HORUS)
• Delay/Jitter Strict (NCSS), Loose (Others)
• Reliability (TCP)
• Path Reliability (OPATM-BFM)
• Amount of Data Exchanged
• MB (FLOAT, GLIDER) Hundreds of GB (HORUS)
• HORUS generates/exchanges massive amount of data

29
Network Services To be Deployed

• Applications that process pre-collected data
• Bandwidth reservation with book-ahead scheduling
• FLOAT, GLIDER, OPATM-BFM, HORUS
• Guaranteed bandwidth connections among SE, CE, IE
  belonging to each VO
• Bandwidth reservation with immediate scheduling
• NCSS
• Guaranteed bandwidth connections among SE, CE, IE
  belonging to the VO

30
Network Services To be Deployed

• Applications that process pre-collected data
• FLOAT, GLIDER, OPATM-BFM
• Premium IP GARR, GRNET
• MPLS VPN GEANT
• HORUS
• Overprovisioning RedIRIS
• Premium IP GRNET
• MPLS VPN GEANT
• NCSS
• Premium IP GARR, GRNET
• MPLS VPN GEANT

31
DORII Applications

• Applications working on data acquired in
  real-time
• Communication Paradigm Point-to-Point
• Communication Mode
• Sensor Network Unidirectional
• VCR, CE, SE, IE, (Web Server) Bidirectional/Symme
  tric
• IPv6, Host Mobility (SMIWR, ODAES)
• QoS Requirements
• Bandwidth 1 Mb/s (EEWS) 100 Mb/s/1 Gb/s
  (ODAES)
• Delay/Jitter Strict Delay (EEWS), Strict Jitter
  (ODAES)
• Reliability (TCP) EEWS
• Path Reliability (ODAES)
• Amount of Data Exchanged
• MB (SMIWR) GB (EEWS) Hundreds of GB (ODAES)
• ODAES generates/exchanges massive amount of data

32
Network Services To be Deployed

• Applications working on data acquired in
  real-time
• Bandwidth reservation with book-ahead scheduling
• SMIWR/ODAES
• Pre-established permanent connections among SE,
  CE, IE belonging to the VO for ODAES is a viable
  solution
• Immediate on-demand scheduling of bandwidth
  guranteed connections
• EEWS
• Guaranteed bandwidth connections among SE, CE, IE
  belonging to the VO
• Pre-established permanent low-bandwidth
  connections among SE, CE, IE belonging to the VO
  for EEWS is a viable solution

33
GEANT2 Services

• IPv6
• Multicast
• MPLS (L2VPN, TE)
• QoS
• Premium IP
• Point-to-Point Dedicated Connections
• VPN
• Security
• Multi-Domain Monitoring
• PERT (Performance Enhancement Response Team)

34
NRENs Network Services
35
DORII Connectivity MAP
36
DORII Connectivity MAP Seismic Applications
NCSS, EEWS
LMU

DFN
USTUTT
PSNC

PIONIER
USTUTT
ECO

INTERNET


GEANT
GRNET

UC
RedIRIS
CSIC
OGS
GARR

CNIT
ELETTRA
MPLS Tunnels
Premium IP/BoD
37
Environmental ApplicationsFLOAT, GLIDER,
OPATM-BFM
USTUTT
LMU
FLOAT


DFN
PIONIER
PSNC



INTERNET
GRNET
ECO
GEANT

UC
RedIRIS
OPATM-BFM
CSIC

CNIT
GARR
ELETTRA

CINECA Server Farm
EUCENTRE
MPLS Tunnels
GLIDER
Premium IP/BoD
38
Environmental ApplicationsHORUS
LMU
USTUTT

DFN

PIONIER
PSNC



INTERNET
GEANT
GRNET
ECO

CSIC
RedIRIS

OGS
CNIT
GARR
EUCENTRE
ELETTRA
MPLS Tunnels
Premium IP/BoD
39
Environmental Application SMIWR
MPLS Tunnels
Premium IP/BoD
40
Experimental Science Application
LMU

DFN
USTUTT
PSNC

PIONIER

ECO
INTERNET


GEANT
GRNET

UC
RedIRIS
CSIC
OGS
GARR

CNIT
EUCENTRE
41
DSA1.2 Table of Contents

• 1. Introduction
• 1.1. Purpose
• 1.2. Application Area
• 2. Network Infrastructure
• 2.1. Network Requirements of DORII applications
• 2.2. Uses case and access networks features
• 2.3. Network Services for DORII applications
• 2.4. DORII Core network deployment over GEANT
  and NRENs
• 3. Network Monitoring Tools
• 3.1. Survey on monitoring tools for the DORII
  network
• 3.2. Deployment of monitoring tools for DORII
  network
• 4. Network Management Tools
• 4.1. Survey on management tools for the DORII
  network
• 4.2. Deployment of management tools for DORII
  network
• 5. Conclusions
• 6. List of Acronyms
• References
• Contact Information

42
Monitoring and Management Tools (1)

• The DORII network
• is a multi-domain one thus, there is the need of
  the deployment of cross-domain monitoring
• contains different network access technologies
  (Gigabit Ethernet, ADSL, WLAN)
• e2e paths have to be established with specific
  QoS requirements
• Network performance monitoring and management
  within DORII will
• Increase network awareness
• facilitate the resolution of end-to-end
  performance problems on paths crossing several
  networks
• e2e troubleshooting is not straightforward
• No standard access to network performance data
  all along the path
• No trust between domains

43
Monitoring and Management Tools (2)

• Provision of performance metrics of specific
  links (throughput, delay, lost packets)
• End-to-end path monitoring that is appropriate
  for Grid aware applications
• Monitoring of the deployed services with DORII
  partners
• Provision of network statistics and tools for
  basic commands (ping, traceroute, etc.)
• Notification messages when service or host
  problems occur and get resolved (via email or
  other user-defined method)

44
Weather Maps MRTG diagrams

• MRTG diagrams that monitor SNMP network devices
  and draw pictures showing how much traffic has
  passed through each interface
• Network weather-maps that display in a visual way
  the utilization of the network links

45
Looking Glass and Smokeping

• Looking Glass (IPv4 and IPv6)
• Smokeping keeps track of your network latency
  (IPv4 and IPv6)

46
PerfSONAR (1)

• PerfSONAR is an infrastructure for network
  performance monitoring that contains a set of
  services delivering performance measurements in a
  federated environment, and it is being used by
  the NRENs in the DORII partners countries
• perfSONARUI can display the link utilization on
  all links of a given path
• An interdomain looking glass makes it possible to
  retrieve, from a single application, information
  from routers located in different administrative
  domains.
• Provides access to already deployed tools
  instances
• Give indication if the problem is rather on the
  network or on the end-site
• Monitor e2e links for specific projects

47
PerfSONAR (2)
48
Nagios (1)

• Nagios can monitor performance metrics of
  specific links and furthermore monitor if
  specific services are online and collect the
  appropriate statistics.
• Monitoring of network services (SMTP, POP3, HTTP,
  NNTP, PING, etc.)
• Simple plugin design that allows users to easily
  develop their own service checks
• Contact notifications when service or host
  problems occur and get resolved (via email,
  pager, or user-defined method)
• Ability to define event handlers to be run during
  service or host events for proactive problem
  resolution
• Optional web interface for viewing current
  network status, notification and problem history,
  log file, etc.

49
Nagios (2)
50
Next Steps

• Define which tools we will deploy
• Define specific e2e paths that have to be
  monitored in addition to DORII network monitoring
• Define alarms for proper network management
• Specific instances may need to be installed in
  DORII partners side (e.g. in PerfSONAR)