Today, it is no longer about the data center.
A digital infrastructure that includes workloads distributed across both physical and cloud environments is becoming the preferred choice for delivering resilient digital services and digital experiences. In fact, according to Gartner, “by 2025, 85% of infrastructure strategies will integrate on-premises, colocation, cloud and edge delivery options, compared with 20% in 2020.”1
However, this enhanced delivery does come at a cost. Enterprises experience greater operational and management complexity, and struggle filling open positions with hard-to-find skills. They also find current data center infrastructure management (DCIM) tools to be woefully inadequate.
The demands of digital infrastructure require a new approach to developing software management tools that can suitably handle the expanded scope and scale. It is for this reason, we designed Hyperview as a cloud native, API-first application that leverages AI to optimize the digital infrastructure.
Diagram: Hyperview is a cloud native application designed to manage digital infrastructures that can include a hybrid mix of data centers, edge facilities, colocation, cloud, and/or IoT.
Cloud native architecture best-fit for big data
Digital infrastructure is, for all intents and purposes, a large enterprise IoT implementation. As such, the nature of the problem is different from 10 years ago when most DCIM applications were designed. The degree of instrumentation has increased, and the rate of infrastructure change has accelerated. This, in turn means that you have to deal with a lot more data and variability within the infrastructure.
Traditional manners in dealing with data ingestion and storage are no longer adequate without significant compromises to the amount of data stored and its granularity. Similarly, from the operational side, the rate at which the application needs to be maintained has changed. Data ingestion can be an issue for larger deployments where a static application is unable to scale and keep up. Whereas, a cloud native application can have its various microservices scale independently and elastically to deal with incoming data. This optimizes cost and puts performance and scale in needed areas of the application when that scale is needed.
As well, most infrastructure and operations teams are averse to managing a large and continuously growing IoT sensor data store. This is not from a lack of ability, but rather from a desire to focus precious skilled resources. A cloud native application built on top of infrastructure engineered to deal with this sizeable data is a better fit than an on-premises application and its inherent limitations. It can be managed at-scale on top of native cloud services for data storage and resilient service orchestration systems like Kubernetes.
Continuous improvement and integration
Putting the volume of data aside, a cloud native application is far better equipped to handle data analysis services, data enrichment, and interactions with other enterprise software-as-a-service (SaaS) products.
Data analysis and machine learning models are continuously evolving. A cloud native application hides the complexity of those continuous updates and tweaks, making it easier for the user to enjoy the benefits. It also makes integrations much more seamless–whether integrating with external data enrichment sources like weather data or SaaS-to-SaaS system integrations through APIs.
AI is a game changer
Gartner predicts that, by 2022, nearly $4 trillion of business value will be influenced by AI to improve the customer service experience, reduce the cost of doing business or generate new revenue. AI is expected to be one of the most disruptive classes of technologies during the next 10 years (see footnote 1).
It is perhaps not a surprise that dealing with a highly instrumented digital infrastructure with thousands of sensors will produce a large data management problem for a standard database. For example, 10,000 sensors being sampled over five minutes will easily produce close to 100 million rows of data in just over a month. Not that long ago, it was acceptable to reduce this data, or to just store the latest values and some rolling averages. However, discarding this data removes any ability for further data analysis, machine learning, correlation/causation statistical analysis, and historical trending. The hidden power of this data becomes very clear as demonstrated by Google DeepMind AI, which reduced Google’s data center cooling bill by 40 percent using data–a sizable feat.
Hyperview’s cloud native architecture gives us the ability to use the right technology to store the data for long-term retention and analysis. It also allows us to use the right tools to analyze and visualize the data in a way that is continuously improving.
Stricter security requires fine-grained access control
Security and access control are critical in any enterprise application, particularly when leveraging cloud native architecture. Other than secure software development best practices, strong identity management and granular, easy-to-manage access control tools are a necessity in managing today’s digital infrastructure. This has become even more evident during the pandemic, where the cloud has become the primary store of identity information. Whether through Okta or Microsoft Azure AD, the key to accessing most applications is by authentication through a cloud service. Flexible configuration and policy toggles within the application should allow for flexible federation with trusted cloud identity stores. A cloud native application like Hyperview leverages this and allows designated authorized enterprise users to access the appropriate information at the right time from any device.
Beyond identity access control, role-based access privileges are another area where system administrators want granular control over what is happening and who can do what. More data and analytics mean more control is needed over what a user can see or do. Very often systems fall into the trap of making this function either too weak or too complex. Striking a balance between function and complexity ensures better overall enterprise security.
1 D. Cappuccio, H. Cecci. (17 September 2020) Gartner: Your Data Center May Not be Dead, but It’s Morphing