Cloud Recommendation Engine

What is the Cloud Recommendation Engine?

Cloud Recommendation Engine is a powerful feature that can provide you with exactly the details you need to plan your next cloud migration, compare costs between AWS and Microsoft Azure clouds, and right-size your next cloud deployment.

Find the Cloud Recommendation Engine (CRE) under Reports > Cloud Recommendation Engine menu, and get clear recommendations for sizing cloud instances to suit your physical or virtual workloads.

CRE feature overview:

  • Cloud Recommendation Engine leverages Device42’s industry-leading auto-discovery and resource utilization data to understand your workloads
  • Cloud instance recommendations are provided for AWS, Microsoft Azure, Google Cloud Platform, and the Oracle cloud platform (including instance sizing recommendations + pricing info)
  • Instance recommendations are calculated based on a combination of your observed workloads (CPU, RAM, HDD, NIC, etc.), utilizing Resource Utilization data if available, PLUS a user-selected “Safety Factor” via the slider.e.g. if your current instance peaks at 4GB RAM usage, and you chose 50% safety factor, CRE will size cloud instance at 6GB:

    D42 Cloud Recommendation Engine Safety Slider

How does Cloud Recommendation Engine Work?

  1. A summary of CPU, memory, and OS info is created based on auto-discovered inventory data
  2. If you have the resource utilization feature enabled, peak CPU and memory usage over the last month is factored in
  3. When you click the “Send to Cloud and Analyze Data” button, Device42 anonymizes your data and sends it to our cloud servers, finding matching AWS or Azure workloads
  4. Device42’s bots do the hard work, returning workload recommendations that have been best-matched each particular device to your main appliance. The anonymized data is then re-matched with your actual device names, and an output sheet is generated that contains both your device names and matched workloads for following 4 scenarios:
    • AWS based on Resource Utilization
    • AWS based on Inventory Data
    • Azure based on Resource Utilization
    • Azure based on Inventory Data

Downloaded Data Sample:
Downloaded Device42 inventory data sample

Google Cloud Platform based on Device42 inventory data sample:
Google cloud platform based on inventory CRE sample

Azure based on Inventory Sample:
Cloud Recommendation Engine Azure based on inventory recommendations

Disk (HDD) and NIC RU details considered by CRE

As of Device42 v16.03.00, CRE also factors NIC and HDD config details as well as new RU information into recommendation calculations [in addition to CPU and Memory (RAM) RU and inventory data as noted above].

The following RU data for NICs and HDDs is now available:

RU-based recommendations:

  • Original NIC Speed (MB/s) & original NIC transfer out (MB)
  • Original Hard Disk Size (GiB)
  • Original Disk IOPS & Throughput (MB/s)
  • Monthly Networking Cost
  • Monthly Storage Cost
Inventory-based (non-RU)

  • Original Hard Disk(GiB)
  • Storage
  • Monthly Storage Cost
NIC-specific data considered:

  • nic_in_speed
  • nic_out_speed
  • nic_in_transfer
  • nic_out_transfer
HDD-specific RU data considered:

  • disk_iops_read & disk_iops_write
  • disk_read_io_rate & disk_write_io_rate
  • disk_read_latency & disk_write_latency
  • disk_total, disk_used, & disk_free

CRE communication details

Cloud recommendation engine communicates with the Device42 main appliance (MA) over https/443, reaching out to either (non-eu) or for EU users. IP address details for this communication (current as of 2018-19) is as follows:

FQDN/IPS:  https/443 https/443