Data Center

Visibility Everywhere in the Data Center Network

by asardell ‎04-02-2017 02:36 PM - edited ‎04-02-2017 02:42 PM (3,251 Views)

Last week we talked about the information we collect from networks, illustrating how network visibility works in concert with automation.  

 

You need to see what's going on at run time in order to take actions. Armed with a closed feedback loop of automation and visibility, you can verify that network updates deliver the behavior you want.        

 

Let’s take a moment to look at the SLX portfolio, with its adaptable infrastructure including inbuilt visibility features. We'll also take an across-the-board look at SLX hardware capabilities, illustrating why the new platforms fit where they do in the modern data center network.

 

Overview of SLX Inbuilt Visibility

 

SLX Insight Architecture and SLX Visibility Services are provided on all SLX platforms.   

 

The SLX Insight Architecture was first introduced earlier this year on the SLX 9850 router. It uses an internal, dedicated 10Gbps path from interface modules to a VM on the management module that supports third-party applications such as TCPDump and Wireshark, and integrates with third-party analytics applications. 

 

Those capabilities are now being extended to the three new SLX switches (SLX 9140, SLX 9240, and SLX 9540) enabling SLX Insight Architecture to be deployed from the leaf all the way to the edge of the data center.

 

SLX Visibility Services are built on programmable silicon capabilities in the SLX 9140 and 9240, complementing the SLX Insight Architecture found across all SLX devices. The “wire to workload” sophistication of the Visibility Services includes:

 

  • Rich (multi-layer) classification and workload matching (Layers 2 through 4) of physical, overlay, and workload visibility at network-wide scale and wire rate performance
  • Automated application of rule-based actions for all of this classification/matching
  • Integration with SLX Insight Architecture, Workflow Composer, and 3rd-party tools

For example, using Workflow Composer, you can react to sensors triggered by a workflow (and using rule-based actions, you can mirror, count or drop selected traffic.   

 

SLX Platform Comparisons 

 

In addition to the visibility described above, the SLX family of switches and routers provide native Support for 25GbE and a future proofing design for long term usage.  

 

The following table gives a cross-portfolio comparison of SLX

 

Table 1: SLX Portfolio Platform Comparisons

 

SLX 9140

SLX 9240

SLX 9540

SLX 9850

Place in Network

ToR

ToR, Spine

ToR, WAN Edge,

Spine, Super Spine, WAN Edge

Interfaces (max)

100 GbE (6)

40 GbE (6)

25 GbE (48->72)

10 GbE (48)

1 GbE (48)

100 GbE (32)

40 GbE (32)

25 GbE (128)

10 GbE (128)

1 GbE (N/A)

100 GbE (6)

40 GbE (6)

25 GbE (24)*

10 GbE (72)

1 GbE (48)

100 GbE (288)

40 GbE (288)

25 GbE (1152)*

10 GbE (1920)

1 GbE (576)

Max I/O Capacity

3.6 Tbps

6.4 Tbps

1.6 Tbps

Up to 230 Tbps

Buffers

24 MB tunable

24 MB tunable

6 GB tunable

Up to 36 GB tunable

ASIC

Cavium Xpliant

Cavium Xpliant

Broadcom Qumran (Dune)

Broadcom Jericho (Dune)

Form Factor

Fixed

Fixed

Fixed

Modular Chassis

(4-slot, 8-slot)

1.5 RU interface modules

 

* Very low distance, not suitable for most applications

 

If an all-Brocade solution is used, all of these platforms are tightly coupled with Workflow Composer so you do not have to do it in the CLI on a box-by basis, you can do it with Workflow Composer. However, you can use SLX Visibility Services with third-party DevOps tools as well.

 

Next-Generation Data Center

 

Now let’s take a look at a complete next-generation spine-leaf data center that scales to thousands of racks and is standards-based for multi-vendor interoperability (Figure 1).

 

slide for slx viz blog.png

 

Figure 1: Visibility and Automaton in the Next-Generation Data Center

 

Use cases include IP Fabric and core-aggregation architectures for both service provider and enterprise data centers.

 

Note: You can also use Brocade MLX series router in core-aggregation core/edge or an IP Fabric Edge. For a VCS Fabric, you would use the VDX Series.

 

Flow Optimizer: Visibility Across the Data Center Portfolio

 

Note that in Figure 1 above, both the Workflow Composer and Flow Optimizer are available everywhere. Flow Optimizer collects sFlow information and sets profiles to view, manage and secure traffic on Brocade ICX, MLX, SLX, and VDX devices.

 

The main use case category is visibility and optimization. For instance, let's say a network with a high-priority video feed is experiencing degradation due to saturation of a preferred link by large IP flows from another source. You can set a policy so that other traffic that is saturating the preferred link is rerouted to other links.

 

The other major use case category for Flow Optimizer is in the area of threat management, dealing with high-volume DDoS attacks at either the infrastructure level (such as DNS spoofing or NTP-based attacks) or at the application (HTTP) level through integration with firewalls and analytics applications.

 

Call to Action

 

For more information on our adaptable infrastructure, visibility solutions, and intelligent automation capabilities, follow the links in this article or contact your Brocade representative.