This syndicated post originally appeared at Network World Zeus Kerravala.
Arista Data ANalyZer (DANZ 2017) provides scalable
end-to-end network and application monitoring
You can’t manage what you can’t see. That phrase that has been used over and over again with network managers as they look to get a better handle on the goings on in their networks.
The problem is pervasive visibility is hard. Collecting information from multiple systems, rolling it up into an aggregated view and then trying to do some kind of manual machine learning on it is next to impossible. I suppose if Mr. Spock were running the network, things might be OK. But the last time I checked, the U.S. wasn’t accepting H1-Bs from Vulcan, so humans will have to do.
The need for visibility has created a rising tide in the network packet broker market, and most of these do a fine job in the enterprise space, but the RISC-based processers that some of them use do not operate at cloud speeds when there are multiple 100 Gig-E connections that need to be tapped. I want to be clear that I’m making distinction between an enterprise monitoring cloud traffic and an actual cloud provider monitoring its internal traffic. The latter is where the challenge is.
This week, Arista joined the many network vendors making announcements by introducing the Arista Data ANalyZer (DANZ 2017) support for its R-Series Universal Leaf and Spine platforms, bringing visibility to 25 Gig-E and 100 Gig-E networks. It’s important to note that the DANZ functionality isn’t a separate infrastructure; rather it’s enabled in Arista’s EOS operating system and combined with CloudVision for automation and telemetry.
The new R Series is a line of fixed or modular leaf-spine switches that can scale from 1 to a massive 115 Tbps. The leaf-spine network can scale up to 432 ports of 100 Gig-E and 1152 ports of 25 Gig-E. As is the case with all Arista products, these are built on merchant silicon—with the innovation being in software—and have the same OS as every other Arista product.
The DANZ feature set can be used for both in-band and out-of-band telemetry and packet capture with a set of features optimized for cloud providers. DANZ features include lossless visibility with deep, flexible buffers; virtual output queue (VOQ) architecture; and microburst detection with LANZ (Latency ANalyZer).
The products can also be used for TAP aggregation through a GUI, CLI or APIs for integration with third parties. The APIs have enabled Arista to build a sizable ecosystem of technology partners that are a who’s who in networking. They include Riverbed, F5, Fortinet, Palo Alto, ServiceNow, Extrahop, A10 and several other best-of-breed companies.
Arista DANZ features
Cloud providers, or even large enterprises with cloud-scale networks, can use DANZ to secure and monitor at link speeds ranging from 25 to 100 Gig-E, enabling customers to do the following:
- Monitor workstreams in real time with CloudVision.
- Automate the visibility of workloads with Network Tracers.
- Mirror in-band traffic with IP tunnels to watch for anomalistic behavior. The traffic can then be capture to the tunnel without additional overhead.
- Mirror to port or CPU to analyze traffic within the infrastructure for doing analysis on the switch. This can significantly speed up problem resolution.
- DANZ TAP aggregation to scale out tools, such as firewalls, load balancers and other network devices.
The DANZ 2017 features will be made available on all Arista R-Series products, as well as the 7150 series of ultra-low latency switches, in the first quarter of 2017. DANZ requires either a monitoring license or a CloudVision license for each platform. No information was given on pricing.
One reason why Arista has been a successful company is that it is not trying to be all things to all people. Customers that need cloud-scale products have benefitted from Arista’s laser-like focus on this market. The DANZ 2017 feature set should be an ideal adjacent market to its core switching products.
Latest posts by Zeus Kerravala (see all)
- Thinking About the Cloud 2020 - March 20, 2017
- SD-WAN: Right for the Cloud Era? - March 14, 2017
- Microsoft and NVIDIA partner to bring GPUs to the public cloud - March 8, 2017