1. Evolution of WIS: Implications from EC-70
Jeremy Tandy, chair TT-eWIS

2. WIS 2.0 – what are the goals?
Give to the users the means to access and use unprecedented amount of data and information from all WMO programmes
Provide easy and reliable access to services (products, tools, applications, computing, storage …)
… while continuing to promote free and open data sharing
Allow users to apply own processing to data
Rationalise common parts of WIS infrastructure
Better serve all types of users, expand the user base, promote interaction between communities: public authorities, research, private, …
source: Matteo Dell’Aqua, chair WMO OPAG-ISS

3. Key goals
Increase the discoverability and accessibility of authoritative weather, water and climate data and information beyond the traditional institutional NMHS user base.
Ensure that the WMO community is equipped to derive benefit from the explosion in data volumes resulting from continued investment in earth-system modelling and observing systems.
Cost reduction for WIS Centres from retirement of the legacy systems and infrastructure, and consolidation of how core WIS functions are provided.
Improve access to data and services in developing countries and, through regional and global cooperation, strengthen the capacity of Members to provide meteorological, hydrological, marine and climate services.

4. Before the discussions at EC-70, our plan was to provide the final draft of the WIS2.0 implementation plan for publication in December
HOWEVER!
EC-70 moved the goal-posts… see Resolution 23 (EC-70) and Decision 18 (EC-70)
The re-drafted Decision 18, plus discussion during the side-meeting had during EC-70 (UK, Japan, Germany, USA, Secretariat with correspondence from Russia), made it clear that we would not be able to achieve approval for WIS 2.0 implementation without detailed technical specification - this was a key requirement from Japan and Germany

5. The finally approved Decision states:
Requests the CBS: […] (3) To provide more information to Members about the technical infrastructure supporting WIS 2.0 and a comparison of the functional architectures for the original WIS and WIS 2.0;
As a result, the submission to Cg-18 would be an update to the WIS 2.0 Strategy and Implementation Approach, plus a PLAN to gather the detailed technical specifications required for Members to assess (and plan and budget for) WIS 2.0 implementation.
Requests the CBS: […] (2) To provide Congress with the updated Strategy and Implementation Approach and a plan to document specific design requirements;

6. Furthermore, implementation approval would be delegated to EC at some subsequent date, allowing for more time to develop those technical specifications
Decides to request from Congress to authorize EC to make a decision on implementation once the updated documents, including design requirements, are submitted by CBS
Secondly, Germany were advocating a decentralised approach to implementation. Reviewing the discussion during the EC-70 side meeting on Cache in the Cloud, and noting our inability to progress the concept of shared services, it seems that we need to fundamentally address the concerns about a decentralised WIS 2.0.
So - to get WIS 2.0 implementation approved, we need to have the technical specifications well understood … so that Members can clearly understand what WIS 2.0 will entail, and how much it will cost so that they can budget effectively.

7. My experience is that good technical specifications ONLY come from implementation experience. So, this means that Members will need to run pilot projects to develop that implementation experience.
Fortunately, Resolution 23 was added (& approved by EC-70) to address this:
Requests the Commission for Basic Systems, in collaboration with the private sector and involving those Members who currently operate GISCs, to analyze the data exchange needs for WIS in order to provide reliable data access to all users considering the availability of new technologies, architectures and communication techniques including cloud based solutions, web services, application programming interfaces, modern messaging protocols etc., and to study how they would support or contribute to the evolution to WIS 2.0 and report back;

8. It also calls for Members, with support from Secretariat, to do the necessary work:
Requests the Secretary-General to provide the resources to support this work;
Urges Members to provide resources for development of potential solutions and feasibility studies.
Thinking out loud (and without detailed insight!), maybe these projects look a little like the GDPFS Severe Weather Forecasting Demonstration Project (SWFDP) [ in terms of trying to implement / demonstrate operational procedures? Projects like these are a large undertaking, and will certainly mean an extension of the timescales for implementation of WIS 2.0.

9. Because we will need to develop the technical specifications BEFORE approval, I wonder whether it makes sense to adopt a more phased approach?
For example, I think that all agreed that the GTS was a major inhibitor to Key Goal A:
Increase the discoverability and accessibility of authoritative weather, water and climate data and information beyond the traditional institutional NMHS user base.
The WIS 2.0 implementation approach document states:
Data exchanged using the GTS in support of WWW are largely invisible or inaccessible to all but institutional NMHS users. Discovery of data through WIS is far from easy; the WIS Catalogue is disproportionately populated by tens of thousands of records describing GTS bulletins, and GISCs lack effective filtering in their search offering. Even if data matching a user’s needs can be found, the majority of users are frustrated by a lack of direct, Web-based access to those data.

10. So phasing out the GTS would be my highest priority (and likely, this would also need to address the 24-hour cache too). The WIS 2.0 implementation approach document describes this as a Core service:
Real-time data, notifications / messaging and a cache of globally distributed data;
We will need large scale pilot projects to determine how to do this; taking into account the need for a decentralised architecture. This should also look at helping Members replace costly private networks with Internet connections (where SLA can be met).

11. WIS 2. 0 also talks about exposing data through Web services
WIS 2.0 also talks about exposing data through Web services. While I think that this is very important, the requirement to push data between WMO Members, particularly observation data, will persist for many years to come.
I also note that there is nothing preventing Web services from being incorporated into WIS 2.0. The challenge is how one provides the metadata that enables those services to be discovered!
Talking of metadata and discovery, WIS 2.0 also talks about leveraging commercial search engines; proposing the use of schema.org structured markup to do so. Again, there is nothing preventing Members who publish their data on the Web from providing this markup and registering those Web pages with search engines. What is required here is guidance on how to do so!

12. If we can address the disproportionate population of the WIS Catalogue "by tens of thousands of records describing GTS bulletins" and encourage publication of data on the Web, then this might improve the user experience sufficiently that we can delay significant work on the WIS Catalogue?
Certainly, I would leave cloud-hosting of services to mature before trying to 'regulate' how this is done. Just like using Web services, there's nothing in WIS Tech Specs preventing a Member from using cloud to meet their needs. Clearly, some level of coordination and sharing of best practice will accelerate adoption.

13. I'll stop hypothesising here. This is for the TT-eWIS team, i. e
I'll stop hypothesising here. This is for the TT-eWIS team, i.e. us, to debate. I'm just illustrating that we could take a different, more pragmatic approach to realising the WIS 2.0 strategy. One that doesn't require such a VAST implementation programme because we focus on delivering smaller tranches of benefit.
So … all of this implies a significant change to the implementation approach! My point? It is
we will need to address concerns around decentralisation; and
that what is submitted to Members in December for consultation is _not_ a final implementation plan for approval. Rather, it will describe the approach in outline and provide a plan for how the detailed specifications will be developed.

14. Resolution 23 (EC-70)
WMO INFORMATION SYSTEM 2.0
THE EXECUTIVE COUNCIL,
Recalling that:
The future WMO Information System (WIS) was first envisaged by the Commission for Basic Systems in 1992, the concept adopted by Congress XIV (2003) and approved by Resolution 2 (Cg-XV), and
The first operational Global Information System Centres (GISCs) began in 2012,
Noting that WIS development, in addition to creating the new functionalities of discovery and access, was confined to be an evolution of the Global Telecommunication System (GTS),
Noting further that:
the current architecture of the WIS was developed almost 12 years ago, and that there are many new technologies, architecture designs and industry solutions that were not available at the time,
Members have made considerable progress in moving from a telecommunications architecture to data centric infrastructure built on the GISCs,
some Members are keen to ensure that WIS 2.0 be developed without such restrictions, making full use of the new information technology and the public/private partnerships environment and practices available to achieve the requirements of Members for the full cycle of information management building on the investments of Members in the initial development and roll-out of WIS,
Requests the Commission for Basic Systems, in collaboration with the private sector and involving those Members who currently operate GISCs, to analyze the data exchange needs for WIS in order to provide reliable data access to all users considering the availability of new technologies, architectures and communication techniques including cloud based solutions, web services, application programming interfaces, modern messaging protocols etc., and to study how they would support or contribute to the evolution to WIS 2.0 and report back;
Requests the Secretary-General to provide the resources to support this work;
Urges Members to provide resources for development of potential solutions and feasibility studies.

15. Decision 18 (EC-70)
The Executive Council noting the WMO Information System 2.0 Strategy, endorsed by Resolution 8 (EC-69) and the Draft WMO Information System 2.0 implementation approach as described in the document presented to the Commission for Basic Systems (CBS) Technical Conference (TECO) (CBS‑TECO‑2018‑Inf‑5(1)‑WIS2‑Implentation‑approach_draft1),
Noting further that CBS Management Group recommended EC-70 consideration for referral to Congress,
Requests the CBS:
To consult with Members on amendments needed to further develop the Strategy and Implementation Approach and specific design requirements;
To provide Congress with the updated Strategy and Implementation Approach and a plan to document specific design requirements;
To provide more information to Members about the technical infrastructure supporting WIS 2.0 and a comparison of the functional architectures for the original WIS and WIS 2.0;
Encourages Members to provide feedback to CBS on amendments needed to further develop the Strategy and Implementation Approach;
Agrees to consider the status of WIS 2.0 at Cg-18 with a view toward implementation and;
Decides to request from Congress to authorize EC to make a decision on implementation once the updated documents, including design requirements, are submitted by CBS.