Presentation is loading. Please wait.

Presentation is loading. Please wait.

WLCG Network Discussion

Similar presentations


Presentation on theme: "WLCG Network Discussion"— Presentation transcript:

1 WLCG Network Discussion
WLCG Workshop 19-23 June 2018

2 Goal: Identify Missing Effort/Capability
We want to conclude the network session with a quick discussion about what may be missing for network monitoring/diagnosis and use compared with identified needs We have heard the experiments provide their perspectives on networking In addition we asked for a 1 page summary of what is needed in network monitoring/diagnosis that follows WLCG Workshop 2017

3 ALICE Concerns Comprehensive WAN network picture is essential for efficient data transfers and placement ALICE depends on and strongly encourages the full implementation of LHCONE at all sites Plus the associated tool (like perfSONAR), which we can use in our Grid frameworks A number of sites in well-connected parts of the world still have to join Entire regions with high local growth of computing resources, in particular Asia, require a lot of further work WLCG Workshop 2017

4 ATLAS Requirements ATLAS would like better visibility into our networks to improve our ability to resolve network issues. Alerting/alarming on network problems: Large packet loss, low throughput, missing data, infrastructure problems. perfSONAR 4.0 campaign Nodes not updating, services not configured properly, services and/or nodes down, problematic test results ATLAS would like to see network traffic data from RENs Dashboards to display relevant data together. Tools to better identify network monitoring infrastructure problems. WLCG Workshop 2017

5 WLCG Workshop 2017

6 CMS Requirements Network requirements scale (to first order) with CPU availability 5MB/s/core would satisfy most demanding workflow – but appears not affordable for large sites Commissioning for 1-2MB/s/core should be the target for LAN capacity A few thousand remote connections of ~0.5MB/s can be expected (for sites with sizable storage) “Full” Tier-2: Many CPUs and large disk capacity Some 10Gbit/s or 100Gbit/s for both LAN and WAN are advisable for sites with several 1000 cores > “CPU-rich” Tier-2: Disk storage for caching only CMS has no experiences with such a site (yet) Mainly the same targets like for the “full” T2 apply > 1-2MB/s/core for LAN (reading from cache) “Good” WAN connectivity (Some 10Gbit/s or more) for filling the cache or reading in (directly) from remote “Disk-rich” Tier-2: More storage than average, perhaps hosting disk for co-located CPU only site Good WAN connection even more important “Disk-poor” Tier-2: Rather similar to “CPU-rich” WLCG Workshop 2017

7 LHCb Monitoring, including network monitoring, available from within Dirac Main use cases: For “helper data processing” sites check the WAN connectivity to a T1 storage For data management check WAN quality WLCG perfSONAR information is included into the DIRAC monitoring infrastructure Consumed via CERN/ActiveMQ Example Dirac perfsonar monitoring: Show packet loss, red if > 5 %

8 GEANT/Mian: Investigating ways to increase ability to differentiate, break dependence on vendors, and provide capabilities to automate / orchestrate multi-domain services Greater differentiation Business and operational systems: digital transformation Service catalogue Service management Order management On-demand request/delivery Service bundling Diversified service portfolio automation / orchestration of multi-domain services Flow-based services VPN services IP services new service offerings Disaggregated control and management plane Multivendor Interoperability Programmable data plane Open & Programmable

9 Questions and Discussion
Need to maintain, monitor and develop our network monitoring; we require people to: Respond to GGUS tickets Track service and metric status and follow up on down services, mis-configuration, etc. Evolve the system, implementing new user interfaces, analytics, alerting and new measurements Coordinate network activities with others Do we have enough people identified to deliver on the experiment expectations? WLCG Workshop 2017

10 Questions and Discussion(2)
Are the services we have (operating and planned) sufficient for identifying and resolving network problems? If not what more is required? Some interest has been expressed about network programmability, especially in light of foreseen, commercially driven network developments. Do we have enough research/prototyping planned to ensure we can utilize future network capabilities? WLCG Workshop 2017

11 References Pre-GDB on networking meeting Summary on Pre-GDB Last LHCONE/LHCOPN meeting WLCG Workshop 2017


Download ppt "WLCG Network Discussion"

Similar presentations


Ads by Google