Valuable HPE insights about the edge for you, from SAP SAPPHIRE 2018

[HPE_Alliances]

The recently concluded SAP SAPPHIRE 2018 event was a major success for HPE and for customers who understand the amazing power and value of HPE – SAP joint solutions. The two excerpts below are excellent examples of the great information communicated at one session on the Edge titled “Big Analog Data – Big Data of the IoT and Intelligent Edge.” They were submitted by one of the presenters, Pradeep Arora.

By Pradeep Arora

Excerpt #1 - The edge must grow as fast data processing moves from the core.

As IT moves into managing OT and IoT, it is important to develop answers that are flexible and modular in their approach. You probably have existing OT technology deployments that are entrenched into your operations. You might think of an end-to-end data pipe line view as a way to help you analyse different options.

I have found it useful to think of response time requirements for the use case I am considering. That gives me functional blocks of IoT, Fast Data, Big Data and AI. This maps logically also to any of the different views and methodologies I have seen in use by my customers.

Figure 1 below illustrates the fact that computing must move to the edge to maximize time value of data and how models must be served as close to the data source as possible.  Thus the reasoning behind the creation of our “HPE Converged EdgeLine Systems”. As HPE’s Dr. Tom puts it “Edge is now the most efficient choice for off prem for IoT and OT”. 

 Figure 1 – An end-to-end data pipeline

Excerpt #2 - Seven things to pay attention to in computing at the edge.

An easy way to think of the edge is where the protocol conversion happens from all the different types of data sources and connectivity to IP-based data packets. Today, you have your data in a computer with an IP address on it. This allows IT to get their hands on it and manage the life cycle.  Edge is the other off-prem away from the core data centre.  Therein lies the challenge for IT teams, whether it be increased surface for security threats, or sporadic connectivity or distributed locations or very low bandwidth and so on.

Pay special attention to 7 things listed in the picture below when you take your applications to the edge.  It’s a totally different world in operations with different SLAs and cost of outage. Look for the round trip latency to where the model will run and design for disconnected autonomous operations. How much will it cost for network bandwidth and cloud computing?

The wired network has been the most reliable one, and will you insist on backend connectivity. Most countries have their own data isolation and data sharing requirements with one of my multinational customers with over 25 VPN in one of their DC. Any store and forward mechanism will have data duplication and the extra efforts required to manage its lifecycle. Finally, many security laws apply to data you might have collected.

Figure 2 -  Seven things to pay attention to in computing at the edge. 

 More information on HPE solutions for the edge is available on the HPE IoT site and on the HPE Edgeline Innovation Network site.