The emergence of the Internet of Things (IoT) has fundamentally changed the way we live, work and communicate. It has also created several business opportunities across segments like smart homes, smart living, digital health and industrial systems. It has been widely estimated by Gartner that there would be more than 20 billion connected devices by 2020 – these heterogeneous devices produce and store huge amounts of data about everything from system performance to consumer behavior. The real business opportunity and technical challenge lies in processing and exploiting this data to create intelligent insights that can provide better-informed business decisions.
In addition, 5G is also set to be a game changer as it brings about a new generation of mobile networks designed to scale and provide blistering data speeds. As is evident with the ubiquity of on-demand video platforms like YouTube*, Netflix* and Hulu*, video traffic continues to grow exponentially and will require significant bandwidth.
However, the convergence of IoT and 5G will require a supporting ecosystem that is tailored to facilitate personalized, smart and secure experiences.
Challenges posed by existing Cloud Infrastructures
Most of the current network architecture design is based on the traditional idea of cloud computing, which is intended to be safe and scalable. All the data from user applications and devices are transported to the cloud and are typically centralized in a data center.
There is the possibility of higher latency as all the data must go to the cloud. For instance, with a video streaming platform, a higher resolution video would require higher network bandwidth and the network delay will drastically impact the user experience. Similarly, high latency can also turn out to be a deal breaker in Industrial IoT applications where safety sensors need to make quick decisions based on the sensitive data that has been captured.
To overcome these issues, a new architecture is required to support services at the edge that can offer the scalability and security of the cloud, and also cope with the additional network complexities.
What is Edge Computing?
Edge computing is a form of cloud computing which brings cloud capabilities closer to the edge. It is the best way to use decentralization to reduce network latency as services that require high bandwidth and low latency can be deployed closer to the user.
Considerations for an effective Edge Computing Architecture
- Virtualization is a key aspect for edge computing to provide agility to deploy the edge applications on demand either as a container or VM. Containers are preferable over VMs as they are more effective at reducing latency.
- Programmable data plane, with high throughput and low latency – Smart NICs.
- Network Slicing, along with QoS implementations.
- Implementing Security for the containerized environment.
- Stateful computing to reduce the latency.
- Reduction of storage latency by localization of the storage node.
Partnering with Intel at the Edge
In building an edge solution, Happiest Minds chose to partner and collaborate with Intel due to its leadership in networking and cloud technologies. This leadership is highlighted in data centers and extends all the way from the core to the edge where Intel brings exceptional computational performance along significant contributions to open standard platforms that offer scalability and flexibility at the edge.
Our edge solutions, depending on the use case, would be built on either Intel Atom® or Intel ® Xeon processors. Security and compression acceleration would be accomplished using Intel® Quick Assist (Intel® QAT), while DPDK would be the foundation for developing the dataplane. OpenNESS, an open source software toolkit to enable easy orchestration and management of edge services across diverse network platform and access technologies in multi-cloud environments is yet another significant contribution from Intel that can accelerate the edge transformation.
Happiest Minds MEC reference implementation based on Intel® Architecture
Application uses Intel® Streaming SIMD Extensions 4.2 (Intel® SS4.2) and Intel ® Advanced Vector Extensions (Intel® AVX) instruction sets to optimize the performance.
Driving Innovation at the EDGE
Multi Access Edge Computing is essentially a cloud-based IT services environment at the edge of the network. It checks all the right boxes by providing ultra-low latency and bandwidth while enabling applications that can benefit from real-time radio network information. The end goal of MEC is to provide an optimized and low latency computing infrastructure with deployment agility that can scale horizontally or vertically based on business requirements.
It goes without saying that the possibilities brought about by the adoption of edge computing are limitless and they create compelling use cases in several areas like industrial IoT, smart homes and digital health.
So, whether it is voice assistants in a smart home being faster and more accurate or whether it is reducing response times for emergencies in smart cities, edge computing is the secret ingredient to drive innovation at the cutting edge.