From digital continuity, to extended enterprise, to the Snowflake Cloud Data Platform (or the other way around)

A while ago I wrote a blog post about the 7 dimensions of the Digital continuity. This blog was about manufacturing and if you look at the common definition of digital continuity aka Digital Thread, it will refer to the ability to seamlessly collect, share and generate accurate insights from data across the product life cycle from its design to its end of life in order to operate the business,

and it's not only applicable to manufacturing but also to all other industries where data is usually produced in silos, has to be processed by multiple specialised solutions, can live is different cloud vendors, must be acquired and shared from/to third parties (internal or external).

Actually if we generalise these topics about seamlessly collect, share and generate accurate insights, we touch the extended enterprise,  "a loosely coupled, self-organizing network of firms that combine their economic output to provide products and services offerings to the market" (*Wikipedia). From a high level perspective, it's a network of players which have been facing multiple data barriers to make the vision happen efficiently.

And to make it happen, it's all about breaking barriers between 

• formats, 
• systems,
• workloads, 
• organisation divisions, 
• regions, 
• customers 
• and suppliers

I recently joined Snowflake with the conviction that Snowflake Cloud Data Platform was able to provide a relevant answer to this vision as a unified data platform able to process any data with any workload at any scale.

After 2 months of trainings and interactions with my peers, partners and clients, I'm more than ever enthusiastic about all the opportunities Snowflake can provide for our clients. The platform provides the ability to break the silos and barriers to unleash new business opportunities by creating a network connecting data from all players seamlessly.

Breaking formats barriers:

Snowflake can load all the data structured or semi-structured. Semi-structured (Json, XML, parquet, ORC, Avro) don't have to be preprocessed to be queried and can contribute to joined queries with other structured data with almost no performance degradation. It means that when building a 360 view you can consolidate data coming from structured sources but also semi-structured such as IoT devices, web logs or social media for example.

Breaking systems barriers:

Snowflake can load all data, whatever the source, using batches or streaming. Thanks to the cloud, there is also no volume limit. It's possible to join multiples sources whatever the volumes and so their structures. By leveraging the VARIANT datatype, it's even possible to adapt seamlessly to the source system schema modification without impacting existing processes as a NoSQL document DB would do (Something I know quite well).

Breaking workloads barriers:

We are used to differentiate data workloads with dedicated solutions and data store (Datalake, DWH, datahub, datamart, etc.). With Snowflake, there is a single version of the data, stored centrally and accessed by multiple workloads concurrently without impacting their performances. Snowflake can instantiate and resize instantly multiple compute clusters with various size to process the centralised data. A single platform to deliver any data centric processing,  from data loading, to data transformation and harmonisation, to data science, to analytics and data applications and all with high scalability and elasticity.

All that is great but here comes the most interesting part especially from a business perspective:

Breaking organisation divisions barriers:

In a typical large organisation, the data is owned by the producers that will eventually share this data with other divisions via large data exports, the data is then transferred, re-ingested, reprocessed, enriched to eventually be combined with other datasets. Considering the energy saving objective to protect our planet, there are some reason to be afraid when we see data duplications and day long reprocessing of something that was existing at the producer in a business ready state.

Now comes Snowflake: with Snowflake it's possible to consume and share datasets with third parties without moving physically the data. The data remains in the data owner account but the third party, the other division in this case, can see and access the dataset as if it was in its own environment. 

It's so possible to consume other divisions datasets and join them with our own. Data is always fresh (as its the one of the owner), doesn't have to be moved or reprocessed before usage. As Snowflake scale independently the compute layer, the third party usage will have no impact on your own performances and the processing cost will also be properly split (the consumer provisions the compute, the owner only provides the storage but as it's already his own data there is no duplication cost).

Snowflake being also available on all major clouds (AWS, Azure, GCP), it also allows to break barriers between divisions who would have chosen different cloud provider, helping to seamlessly perform global consolidation. The data can indeed be shared between cloud providers with a unified platform.

Breaking customers and suppliers barriers:

Actually what we just described for divisions barriers is also applicable for customers and suppliers. Snowflake allows to share data with and from third parties. These third parties can be uniquely identified and authorised (sharing stock inventories, product data, sales, data, etc)  or can also be members of the data marketplace where data provider can share and eventually monetize their datasets.

One more time, the data is stored once in the owner account and can be accessed and joined with consumer datasets without having to transfer it with large batches and processing.

All these capabilities provides the foundations to create new use cases by breaking the barriers we all know so well.

Hope this provides you some insights about my enthusiasm for Snowflake. I'll soon publish more industry focus content. If you want to know more don't hesitate contact me or follow me on LinkedIn.