As IoT devices and applications continue to proliferate, the centralized, cloud-based model used by many enterprises is taking a back seat in favor of moving data and processes closer to the edge to increase their flexibility, agility, and accessibility.
With more and more organizations adopting real-time capabilities and 5G networks on the rise, 2021 is poised to be a banner year for digital transformation opportunities. However, challenges remain, from the well-tread issues of speed and scalability, to data security and governance.
DBTA recently held a roundtable webinar featuring Lewis Carr, senior director, product marketing, Actian; Mark Gamble, product and solutions marketing director, Couchbase; and Shane Johnson senior director, product marketing, MariaDB, who discussed key technology, architecture considerations and emerging IoT best practices.
Edge computing is a distributed computing framework that brings data and compute closer to the applications that consume it, Gamble explained.
Edge computing offers:
· Reliability and resilience: Uninterrupted access to data regardless of Internet connectivity. Improve customer experience and improve guarantees on business uptime.
· Speed: Guaranteed low latency and sub-millisecond data access regardless of Internet bandwidth. Essential for near-real time, mission critical applications.
· Security: Data privacy, data governance, and regulatory policy compliance - sensitive data never leaves the edge.
· Bandwidth: Save on network bandwidth costs and reduce the load on cloud servers by reducing volume of data transferred over the Internet.
There are workload differences between traditional SQL data and IoT, Johnson said. Traditional SQL data tends to be mixed read/write, optimize read performance, and almost always transactional. Whereas IoT can be write intensive, is optimized for write performance, and is best for hybrid transactions/analytics.
There are scalability challenges from traditional to IoT that includes:
· Going from 100s to 1000s of TPS, to 100s of thousands or millions
· Going from 100s to 1000s of concurrent customers, to millions of devices
· Scaling reads is not too difficult, but writes…
According to Johnson, the perfect solution is MariaDB Xpand. The platform offers unlimited scalability, data integrity, and high availability. The MariaDB platform is optimized for elasticity and scalable transaction processing with distributed SQL-scale out to execute millions of transactions per second with low latency and high availability, and to support very large databases with billions of rows.
According to Carr, Gartner analysts are saying:
· By 2022, Organizations utilizing active metadata to dynamically connect, optimize and automate data integration processes will reduce time to data delivery by 30%
· By 2022, one-third of complex field service organizations will utilize machine learning to predict work duration and/or parts requirements, rising from less than 2% today
· By 2023, customers will prefer to use speech interfaces to initiate 70% of self-service customer interactions, rising from 40% today
· By 2023, at least 35% of midsize to large enterprises will leverage a hybrid cloud-to-edge computing deployment model for at least one IoT project
Carr said cloud will only be part of the solution based on three reasons including latency, security and data privacy, and innovation at the edge.
Carr explained that “50% of organizations that are subject to GDPR will suffer from a lack of governance and visibility into mobile workflows.” By 2022, IoT security attacks due to lack of insight into edge and third-party device providers will increase by 35%.
From a data management perspective, Carr suggests encrypt at rest and in transit to reduce attack surface. Remove ETL between devices and tiers to eliminate attack vectors. Process at the edge, mask results, and eliminate raw data to improve privacy. And use industry standard DB technology and its inherent security to support better data governance at the edge.
For innovation at the edge, according to Carr, break down proprietary barriers and walled gardens. Create a vibrant developer base by courting Enterprise and Mobile application developer communities. Anchor a store mall approach to building ecosystems. Identify and successfully pilot key Edge use cases.
An archived on-demand replay of this webinar is available here.