Edge computing requires a change in performance monitoring
Our smart, connected devices can generate and processing more data than ever before. The mountain of available data fuels valuable new services and applications, delivered to users through a combination of traditional, cloud, edge, and fog computing technologies. As a result, the market for edge computing is exploding at an estimated CAGR of 34% through 2024, according to Reportbuyer, with adoption led by the Banking and Financial Services sector. [1]
As users, we expect service delivery to be fast, location-aware, uninterrupted, and available wherever we are. To satisfy these demands, many enterprises and government entities are moving applications and data processing to the edge of their networks, closer to their users.
This shift to edge computing requires a complementary shift in how we monitor application performance and user experience. Transferring large volumes of traffic to the data center to process with enterprise-scale Application Performance Monitoring (APM) and Network Performance Monitoring (NPM) solutions is costly and takes time. Deploying enterprise-scale solutions at remote sites and branch offices is expensive. Smart enterprises address these challenges by incorporating edge monitoring: collecting traffic where it is generated and sending it directly to cloud-based APM and NPM solutions--bypassing the data center. Edge monitoring helps IT identify and resolve performance issues more quickly and can reduce operating costs.
Traffic capture on the edge
The first step is to gather all the edge traffic you need to monitor. Network packet brokers (NPBs) are the industry’s leading solution for aggregating and filtering network packets for security and performance monitoring. Traditionally, companies deploy NPBs as powerful, purpose-built hardware devices in the data center. That works well when critical applications run in the data center and traffic is transferred there anyway. In edge computing, however, some transactions will not flow through the data center at all. To monitor traffic at the network edge, companies need NPBs packaged and priced appropriately for use at remote sites and branch offices. Instead of a few regional data centers, this could mean hundreds or even thousands of satellite locations. You still may want to transfer selected traffic to the data center for advanced forensics and correlation analysis, but that will be a subset of the total traffic. For routine, high-volume monitoring, it will be more efficient to process traffic right at the edge.
To support this shift, Keysight has released edge-specific versions of their industry-leading Vision network packet brokers (NPBs). The Vision Edge 10S and Vision Edge 1S are sized to handle the monitoring needs of a single rack, branch office, or remote site. These solutions perform:
- packet aggregation and replication,
- packet filtering,
- VLAN tagging and untagging,
- deduplication,
- packet trimming and header stripping,
- GRE tunneling, and
- load balancing
They can be rapidly deployed and modified quickly with Keysight’s NPB easy-to-use web interface and are affordably priced for branch office/remote site deployment.
In addition, the Vision E1S NPB comes pre-installed with Keysight Hawkeye network performance monitoring software. Hawkeye lets you automate performance checks to monitor the experience of end users. This works well at sites with no local IT personnel since Hawkeye is configured and managed using a cloud-based interface and dashboard. Hawkeye also lets you proactively probe your network for vulnerabilities by processing synthetic traffic through your network and comparing metrics to predefined key performance indicators (KPIs).
Other cost saving options
Even with an attractively priced Network Packet Broker (NPB), it can be expensive to enable traffic filtering at hundreds of remote sites. Two other cost-saving options are also available.
The first option is deploying Keysight Vision Edge packet broker software on an open switch. With this approach, organizations use their production switches as NPBs, leveraging existing equipment investments and reducing implementation costs. Open switches running Vision Edge OS seamlessly integrate with Keysight NPBs in the data center for a complete enterprise solution. Learn more at Vision Edge OS | Keysight
Another cost-saving option is available for users of Cisco Catalyst switches. Keysight and Cisco have teamed up to let customers install Keysight Hawkeye network monitoring software directly on Catalyst 9000 switches using the Cisco DNA Center management platform. This approach uses widely deployed Cisco hardware and eliminates the need to purchase additional equipment. DNA Center lets you download Hawkeye automatically to all your Catalyst switches for fast and error-free deployment.
Summary
Edge computing is becoming a strategic choice for optimizing edge performance and user experience. As you move applications to the edge, evaluate your options for monitoring performance there as well. The cost of implementing edge-oriented network packet brokers may not be as much as you think. You also may be able to leverage existing switches to run network visibility and performance monitoring solutions.
[1] “Edge Computing Market – Growth, Trends and Forecast (2019-2024),” PRNewswire.com, June 21, 2019.