We continue the series of materials about Huawei networking products with an overview of CloudFabric Easy DC solution: how it differs from the "mother" CloudFabric, what advantages it offers and on which hardware it is built. Plus, how to create a detailed design of your data center and get a package of its specifications simply by filling out the online form in the Huawei Smart configuration tool.
Often when designing small data centers, designers are forced to take networking equipment and platforms with disproportionate performance. It is great to have a head start for the future, of course, but sometimes this head start is not economically feasible. If not, it may not be used at all.
That is why we have made a network "factory" specifically for small data centers. Including those where the standard three-tier network infrastructure model is redundant.
General view of CloudFabric
So how is CloudFabric different from CloudFabric Easy DC? First and foremost, in scale. But before getting into details, let’s run through what CloudFabric is all about.
To make a long story short, it is "factory" by Huawei It was introduced to the market a few years ago as a response to market trends: CLI (command line interface) gradually gives way to API (application programming interface); controllers appear to solve the tasks of changing the configuration of active network equipment; the number of virtual entities (virtual machines, switches, routers, ITU) multiplies; the number of threads increases, their width increases. As a consequence, traditional monitoring and management tools are no longer sufficient.
As network facilities have improved, the issue of full-fledged flow analysis (collecting and evaluating telemetry data) has become acute. This can be done using various network protocols, accumulating statistics of the operation of network equipment online. The captured information is packaged, and this telemetry data is sent for processing to the collectors of the flow telemetry analysis system. The result is a versatile platform that can handle three main tasks:
- fast automated troubleshooting;
- ability to predict service states, including for certain configuration changes;
- ensuring information security due to the fact that the system is able to independently identify a number of threats that are not fixed by modern means of intrusion detection.
Another notable trend is the proliferation of "factories" consisting of underlay networks (standard protocols and fixed configurations) and overlay (service network for the customer). Finally, a modern networking system should be open, integrating through "north" and "south" interfaces.
We’ve implemented all of the above ideas and tools into CloudFabric. Its functionality includes underlay and overlay network deployment, resource orchestration, service chain organization. Last but not least, network diagnostics and TTRs (Time to Repair) are handled by our Fabric Insight telemetry analysis system.
CloudFabric includes iMaster NCE Fabric, a network monitoring and management system that is configured to select the features needed for a specific project: SDN Controller (AC-DCN), analytical tools (Fabric Insight), firewall policy management and configuration synchronization tools (SecoManager), and standard monitoring tools (eSight).
Now back to the CloudFabric Easy DC platform. This is a slightly lighter version of CloudFabric, but it includes an additional toolkit that allows you to quickly build the right specification.
Positioning CloudFabric Easy DC
As mentioned, the CloudFabric Easy DC platform can handle both underlay and overlay networks as well as integrate with overlay platforms. It supports microsegmentation and Service Function Chaining, has logical network visualization tools and has the ability to add Value-Added Services (VAS). Fabric Insight functionality is available as an option.
CloudFabric Easy DC has limitations on the number of active network devices and computing platforms. Let’s say Fabric Insight, as standard, uses several data-collecting systems, several analytics compute nodes, and a storage platform. That’s five or six servers minimum. In the CloudFabric Easy DC solution, the Fabric Insight platform is represented by a single server. Hence the other limits :
- max 600 servers;
- of virtual machines or containers up to a maximum of 24k;
- no more than two spines (spine switch);
- no more than two border leaf switches, etc. (see figure above).
All of the above makes CloudFabric Easy DC the optimal choice for rapid design of HLD solutions and running small data centers (under a hundred servers).
Since we only have one server without separate collectors in this solution, the full functionality of flow analysis is not available, but all other telemetry tools (CPU and memory usage, MAC- / ARP- / FIB-tables download, interface and queue statistics, packet loss statistics and optical transceiver info) are still available.
Sample HLD solution based on CloudFabric Easy DC
Let’s take a typical solution of a small data center with :
- computing load (physical and virtual machines);
- network infrastructure;
- Internet and WAN outputs.
CloudFabric Easy DC allows you to cover several similar data centers – with an eye to the limitations mentioned in the previous section.
The diagram shows the usual three-tier architecture: server leaf connected to the computational load, spine network core and border leaf connected to both external and service networks as well as to the neighboring data centers. The hardware component of the core is represented by CloudEngine CE8850-64CQ-EI switch, at the access layer – CloudEngine 6865-48S8CQ, but if you need an electrical interface, it is reasonable to use CloudEngine 6856-48T6Q-HI. Huawei is positioning this infrastructure solution as a direct competitor to the Cisco mini APIC.
When three-tier architecture is unnecessary
While three-tier architecture is common, in some cases it is not optimal. You can start building a data center with just two switches that have all the necessary functionality, and then grow horizontally, introducing spine and border leaf.
iMaster NCE as a separate DCN analyzer
The iMaster NCE, complete with Fabric Insight, can serve as a standalone DCN analyzer and collect data in a switch-based network. It’s a true easy solution that leverages our signature 1-3-5 approach to network troubleshooting: find the problem in one minute, fix it in three, restore service in five.
Simple calculation of data center design
Gone are the days when you had to work in Excel calculating transceivers and ports by the number of uplinks. Now, all these and many other network building activities can be performed on the automated web-based Huawei Smart configuration tool.
In the first step (Requirement design), the designer, filling out the online form, selects a suitable design, determines the availability of optical or electrical ports, specifies the number of data centers.
Next, you can select the iMaster NCE mode of operation and the required types of x86-based servers. Among other things, the servers can be provided by the customer or deployed in a virtualization environment altogether. Both our Huawei FusionSphere and Microsoft and VMware solutions are available as management software.
During the Product Sourcing phase, the designer has the opportunity to select the types of power supplies, determine the number of leaf switches and transceivers in the uplinks, indicate the need to deploy microsegmentation services and the Fabric Insight platform. Communication channels with neighboring data centers are also configured separately. The Smart configuration tool constantly checks the design for errors and, if there are any, reports them.
Once the configuration is complete, the Smart configuration tool performs a design calculation and provides an uploadable set of HLD documentation ready to be sent to the customer. The report includes the specifications of each device, network topology information, traffic patterns, configuration recommendations, cost calculations, and more.
To learn more about Smart configuration tool and to see the configurator in action, please watch our webinar or directly access this tool on the Huawei website (if you have partner access rights).
Benefits of CloudFabric Easy DC
For comparison, let’s take the previously mentioned Cisco mini APIC. In the standard architecture (two spine and two leaf) NCE functionality (without Fabric Insight) can be implemented on virtual machines. This gives a rack space advantage over a competitor’s solution.
If the data center is going to grow, it is reasonable to use extended design and physical servers. In particular, to be able to organize a Multi-PoD design, when applied, the "factory" is able to run on multiple physical data centers. Clustering of physical servers allows not to depend on the failure of one of the components, as well as to use clusters "stretched" between two data centers (if delays allow).
High availability solutions based on CloudFabric Easy DC enable system uptime if any of the three servers in the cluster fails.
Finally, the technical characteristics of the CloudFabric Easy DC switches make it possible to build solutions with higher performance than the industry’s closest to them.
Learn more about switch models
The CE8850-64CQ-EI 2U core switch is built on a Broadcom chipset and supports telemetry capture via gRPC and ERSPAN, which can then be transferred to Fabric Insight. By collecting TCP packet headers an administrator can fully analyze any session. As for UDP traffic, it requires a special AI Fabric chip, which is also available in some of our solutions.
The CloudEngine 6857 platform is equipped with 10Gbps interfaces. The telemetry capabilities of the 6857 are exactly the same as the previous switch, all other key features are noted in the diagram.
Finally, the CloudEngine 6863 switch is available for order with electrical interfaces.
We would like to add that the line of Huawei network equipment continues to move from third-party chipsets to its own, expanding its capabilities. We are waiting for switches which are built on our chipset and which fully cover all the functional niches in the data center segment.