As a kid, I used to love to look at the sky and see what forms I could find in the clouds. Still do.
But as a professional adult, in order to guide customers to achieve business outcomes from their analytical ecosystem, I need to understand what is happening in the analytics industry, map what trends are coming, and enable customers to get the best from modern architectures in the cloud.
As part of this ongoing process -- distilling and understanding everything related to analytics in the cloud –- I've recognized some critical capabilities in modern cloud analytic architectures that your organization should demand, no matter what.
Not just multi, but hybrid cloud
A multi-cloud vision in your data & architecture strategy will guide you to get the best of the cloud, with no lock-in and high levels of availability. But it’s important to respect and understand the data gravity that may drive you to a hybrid cloud approach -- mixing analytics on the cloud and on premises. Companies will want to protect their decades of investment but will also require innovation as part of their digital transformation, so it is necessary to consider technologies that can be portable with no recoding.
Be modular, separating compute and storage
Elastic scalability is what provides the flexibility that users are looking seeking when the move to the cloud. Being more cost-effective, while also enabling self-service analytics, puts you in better shape to achieve your analytic goals in the cloud. It’s not just about enabling scale in/out, up/down, start/stop controls, but letting users manage them without asking someone else.
Integrate with cloud services provided by cloud providers
It’s not having a standalone platform that matters, but how your ecosystem is architected to get the best from that analytics platform, combined with cloud services, in a multi-cloud fashion. Leverage those services across Amazon Web Services, Microsoft Azure, and Google Cloud Platform. Some examples are Amazon EBS, Glue, Lambda, SageMaker, S3 and Azure Blob Storage, Databricks, Data Factory, ML Studio, Power BI.
It’s not just about structured data
New data formats are popping up from different sources around your organization. A modern analytics platform must process not just structured data but newer semi-structured and unstructured data types. Your analytics initiatives will improve and deliver better results when considering all the new data coming from data ingest pipelines such as JSON, Avro, Parquet, XML.
Be inclusive, there are new kids in town
It’s not realistic to force users to consider other languages and tools they’re perhaps not familiar with. Be open! Modern platforms must be flexible to support the handling of complex algorithms and be scalable to allow for traditional and newer analytical techniques (machine learning, deep analytics, artificial intelligence).
No, you can’t live without a dynamic resource allocation capability
A modern analytics platform must have a dynamic resource allocation capability to allow support for different analytical workloads that users will submit to the ecosystem. This is not an option. Even with elastic scalability capabilities, time-to-market requirements from business users will require you to guarantee service level agreements. Give them your commitment to achieve those SLAs.
You will not find all these critical capabilities together in any modern architecture, but Teradata Vantage
can deliver them all. Vantage is not just the best SQL engine for analytics. It has evolved to solve a brave new world of data & analytic challenges. Vantage provides a flexible dynamic resource allocation capability, allows for the new languages and tools your users are leveraging, supports new data types that new applications and systems are delivering, integrates cool services provided in the cloud, scales elastically thanks to separating of compute and storage, and can be deployed anywhere – in the cloud, on multiple clouds, on premises, or any combination thereof.
Welcome to the future.