Foster innovation with Dell APEX Cloud Platform (ACP) for Microsoft Azure with Dell PowerSwitch
Modern enterprises depend on their IT departments to facilitate their intricate operations and maintain a competitive edge in the market. To secure their competitiveness and achieve success, these companies aspire to develop groundbreaking applications and services with remarkable agility. Many Businesses are embracing a multicloud approach to fuel their inventive endeavors and expedite the deployment of applications.
Dell APEX Cloud Platform for Microsoft Azure is a hyperconverged infrastructure (HCI) solution that is engineered collaboratively with Microsoft. The solution optimizes Azure hybrid cloud operations which consolidates compute, storage, and network into a single, highly available, unified system. With proper planning, ACP for Azure rapidly deploys into an existing data center environment, and the end-product is immediately available to deploy applications and services.
Figure 1 APEX Cloud Platform for Microsoft Azure.
Dell Networking for Dell APEX Cloud Platform for Azure
The network considerations for Dell APEX Cloud Platform for Azure are no different than any enterprise IT infrastructure: Availability, Performance, and Scalability. Dell Technologies manufactures APEX Cloud Platforms (MC760 and MC660) for Azure in the factory, per customer’s purchase order and delivers to the customer’s data center, ready for deployment. Dell has tested the overall solution with Dell PowerSwitch platforms. The nodes in the ACP for Azure attach to Dell’s Top of Rack (ToR) networking switch which meets the Microsoft Azure Stack HCI network functional requirements.
Following are the prime requirements for networking:
- Priority Flow Control (PFC) – You require PFC where you use Data Center Bridging (DCB). Since you can use DCB in both iWARP Remote Direct Memory Access (RDMA) and RDMA over Converged Ethernet (RoCE) scenarios, you need PFC in all scenarios.
- Enhanced Transmission Select (ETS) – You need ETS where you use DCB. Since you can use DCB in both RoCE and iWARP RDMA scenarios, you require ETS in all scenarios.
- Link Layer Discovery Protocol (LLDP) – You require LLDP as it enables troubleshooting of physical networking configurations. You must dynamically enable configuration of the LLDP Type‑Length‑Values (TLVs).
- The required organizationally specific custom TLV (TLV Type 127) subtypes are:
- Port VLAN ID (Subtype 1)
- VLAN Name (Subtype 3) – Minimum 10 VLANs support
- Link Aggregation (Subtype 7)
- ETS Configuration (Subtype 9)
- ETS Recommendation (Subtype A)
- PFC Configuration (Subtype B)
- Maximum Frame Size (Subtype 4)
- For Software-Defined Networking (SDN) encapsulation, you require a Maximum Transmission Unit (MTU) range of 1514 to 9174.
- You require Border Gateway Protocol (BGP) to enable tenant workloads with SDN and dynamic peering.
- Also, you need a Dynamic Host Configuration Protocol (DHCP) relay agent for Preboot Execution Environment (PXE) boot services.
Network redundancy and performance considerations
ACP for Azure is a software-defined data center which depends on the physical top-of-rack switching for network communications and is engineered to enable full redundancy, plus failure protection across the cluster. For customer environments that require protection from a single point of failure, you must design and configure the adjacent network supporting the ACP for Azure cluster to eliminate any single point of failure. A site must deploy a minimum of two Dell Top-of-Rack PowerSwitches to support high availability, plus balance the workload on the ACP for Azure cluster. A pair of cables (VLT Connection) links these Dell PowerSwitches which supports the flow of Layer 2 traffic between the ToR switches.
Scalable ACP networking topologies for Azure
Customers can implement various networking topologies to expand to a maximum cluster size of 16 nodes. Dell offers the following scalable networking topologies for APEX Cloud Platform for Azure:
- Fully converged – RDMA, cluster management, and VM traffic traverse the same Ethernet connections thus conserving on switch ports and cabling requirements per node.
- Non-converged – Separates RDMA and host management/VM traffic on to separate network adapter interfaces. This ensures that there is no contention between storage and LAN communications, plus is easier to troubleshoot.
Figure 2 Dell’s scalable ACP network topologies for Azure example.
Dell-on-Dell value proposition
Having an end-to-end stack from Dell Technologies enables customers to build a cohesive and efficient IT infrastructure. This allows customers to focus on their core business objectives rather than managing complex and disparate infrastructure components. Dell-on-Dell’s value proposition for integrated networking, storage, and compute solutions offers the following benefits:
- Seamless integration of Dell networking with Dell APEX Cloud Platform for Azure which simplifies deployment, management, and maintenance, plus reduces the risk of interoperability issues.
- Dell APEX Cloud Platform for Azure, deployed with Dell Networking, optimizes and overall provides better system performance.
- Single point of support across overall deployment, provides a consistent service experience.
- Dell APEX Cloud Platform for Azure solution, with Dell networking, offers competitive pricing compared to stand-alone components from various vendors.
- Reduced complexity and efficient management translates into lower operational expenses (OPEX).
- Regular and seamless system updates across the ACP for Azure ecosystem.