These decisions require a company to weigh the cost of edge compute versus the cost of data transport. Cost can come in different forms; dollars are an obvious cost, but time is another cost factor — and attempting to transport large amounts data for analysis on a time-critical function is not realistic and requires more localized compute capabilities. Thus, the need to optimize server selection based on your edge use case.
Choosing a server for your edge application is certainly based on the application needs, but the physical and environmental constraints your “edge” requires are also significant factors. In addition, the lifecycle management and service model of the infrastructure, as well as the resiliency expectations of the application are also key to choosing the type of infrastructure to manage your edge’s data lifecycle. Factors around workload requirements, environmental constraints and lifecycle management create a laundry list of questions for customers to address. Below we introduce three scenarios illustrating the changing and varied edges, and a few ideas on how to choose the right server for your edge.
Use case #1 – The Retail Edge
Smart retailers are changing their tactics from mass marketing to personalized experiences and from mass advertising to omnichannel customer outcomes. The omnichannel shopper is the customer who uses multiple touch points before making a purchase through web, voice, digital, email and in-store shopping. In the future, those disparate activities will become smarter and will have the ability to share information between platforms, providing a seamless experience. Today, you no longer get demographically generated coupons from cutting edge retailers. Instead, you receive coupons for items you have historically used and are interested and intend to buy. This type of attention to the individual generates brand loyalty and a differentiated customer experience – something all retailers are striving to achieve.
This has created a focused effort to understand compute, storage and communication requirements for the Retail Edge to provide a differentiated shopping experience. Shopping areas, while hospitable for casual perusing and optimized to move merchandise are potentially challenging compute environments. The server whose design depends on stable power, cooling and ample service space in a traditional data center will not be optimized in a retail environment. Reconciling traditional server designs and retail edge requirements takes a little effort and time to address.
Although there are site requirements and constraints to consider, typically the larger challenge with the retail edge is around infrastructure management. Since the skillset is typically not IT focused, there is a need for remote manageability using the same control surface administrators use today when managing data center servers. Understanding this, Dell EMC has a set of remote administration tools using the latest version of Dell Remote Access Controller, iDRAC 9 Datacenter, where customers can remotely process streaming data analytics critical for understanding edge operations to meet their requirements for deploying, securing and operating edge environments. With the added telemetry streaming on iDRAC9, customers can discover trends, fine tune operations, and create predictive analytics to help ensure peak performance, reduce downtime and prevent risk. When integrated with other Dell EMC products, like VxRail, customers receive a solution that consolidates compute, storage and virtualization with end-to-end automated lifecycle management and simplifies the edge management experience.
Use Case #2 – The Industrial Edge
Data has been generated at the “industrial edge” for years, but the analysis of the data onsite was challenged. First, to perform the proper analysis you needed the appropriate level of compute and storage resources onsite. As an example, for years offshore drilling companies would carry volumes of drilling data stored on tape to a land-based data center via helicopter because the cost of localized compute and storage to process data in a relevant time frame outweighed the cost of data transport. The other problem was related to maintaining cost effective IT equipment in very harsh environments. While compute effectiveness has increased, the harsh conditions remain. This is obviously not isolated to the oil & gas industry, as we see similar challenges for factories where an increased amount of compute is needed to sustain operation in an environment requiring unique filtration for equipment, and also for military applications where data center environments need to be quickly established and mobile.
Although standard data center equipment could otherwise be used, the environmental constraints of the industrial edge create the need for a level of ruggedness that could put a cost premium for onsite components. Instead of adding this premium to each IT component, customers now have the option to leverage Dell EMC’s Modular Data Center (MDC) and use a containment solution that allows standard data center equipment that is cost-optimized and readily available to be placed in a secure, ruggedized and manageable enclosure. Whether the need is for a few racks of equipment or for only a few servers, an MDC allows you to quickly establish an onsite, managed facility for the necessary compute and storage resources for your edge environment. A couple of specific products that address these needs are the MDC Micro 815 which provides up to 48 rack units of space and up to 30KW of power, and the MDC Micro 415 with 17 rack units of space and up to 8KW of power. These MDCs can provide customized capabilities for mobility, security, filtration, cooling/heating, ruggedness and power so standard data center equipment can be used in harsh conditions. This allows customers to choose any PowerEdge server based on workload requirements without being constrained by environmental conditions.
Use Case #3 – The Automotive Edge
It’s no secret the automotive industry is going through an IT led transformation. These changes include more compute, storage and networking capabilities not only in the vehicle, but at every stage of the data pipeline from the car, to the core, to the cloud. In-car compute systems will need to be coordinated with curbside aggregation devices with miniature roadside data centers performing a myriad of simultaneous operations for thousands of vehicles like computation, analytics, filtering and storage choices; choices that can only be accomplished when compute and storage are close to where data is created and consumed. Layering autonomous vehicle capabilities on top of the already connected ecosystem will dramatically increase sensor data from Radar, GPS, LIDAR, Ultrasound, and Video subsystems. Including infotainment options and the list of the right kind of compute, storage and communication in all the right places throughout the entire edge environment becomes vital to increase the amount of awareness, coordination, synchronization and movement of data.
Educated customers don’t want to buy a singular hardware offering, they want to buy hardware that is integrated into a larger system that enables a uniform end-to-end experience. Right-sizing a server to fit into a larger edge environment requires a good understanding of not only the workloads, but the location and the physical environment, how data will be lifecycled and how the end-to-end management is enabled. The automotive edge encapsulates several of the requirements and considerations from the retail edge and industrial edge use cases. Servers must have I/O flexibility, varied processor offerings, provide distributed management, work with constrained power, operate in potentially harsh conditions and integrate into a broader solution.
These diverse edge requirements are addressed with the recently announced PowerEdge XE2420 and Dell EMC Streaming Data Platform. The XE2420 is a short-depth, hardened hardware platform with the latest iDRAC system management and telemetry streaming that optimizes system latency, supports significant storage and provides support for network and acceleration flexibility. When this server is used with the Dell EMC Streaming Data Platform, customers can ingest streaming data and unify data regardless of source or type for analysis, giving customers the ability to generate insights and extract business value from edge applications.
As you can see, choosing a specific server for one edge does not fit all. From a hardware resource and capability, design ruggedness or a lifecycle management model each edge has its own set of requirements. Customers must understand their workload requirements, their environmental constraints and their “cost” of data compute at the edge versus the cost of data transport to the core or cloud. This is precisely why it’s important to have a solid portfolio of choices for customers and a consistent, scalable management framework across all those portfolio choices. Dell EMC has achieved the #1 server vendor position by working with customers to understand these varied requirements and how best to address those requirements.
When you buy a PowerEdge server, you’re not just buying a point product, you’re buying into a world-wide supported, extended ecosystem the enterprise administrators are already familiar that gives them a consistent control plane from edge to core to cloud. Additionally, the broader set of Dell Technologies products provide a set of offerings to store, transport, analyze and secure data across the edge data management lifecycle. As the edge matures, Dell EMC will be the right partner to help you choose the right server for your edge.
0 comments:
Post a Comment