Flexible Engine Elastic Cloud Server introduces new D3 instances for disk intensive use cases
Flexible Engine
Release Notes
July 1, 2020
Elastic Cloud Server (ECS) is Flexible Engine’s virtual machine service based on Openstack which offers a wide choice of flavors designed to run your workloads on highly performant instances at the best price.
ECS has received the following updates :
New feature : Default AZ placement for the instances of your tenant
Each region of Flexible Engine is designed with 3 Availability Zones (AZ). The region and availability zone (AZ) identify the location of a data center. You can create resources in a specific region and AZ. This is the case for ECS instances, for which you have to choose a region and an AZ during the creation process. Before describing the new default placement feature, here are a few characteristics of regions and AZ:
- A region is a physical data center, which is completely isolated to improve fault tolerance and stability. The region that is selected during resource creation cannot be changed after the resource is created.
- An AZ is a physical location where resources use independent power supplies and networks. AZs are physically isolated but interconnected through internal networks. Because AZs are isolated from each other, any fault that occurs in one AZ will not affect others.
Each time the AZ needs to be chosen inside the region at the creation of a new ECS instance. We are now introducing a new default placement feature which allocates always the same AZ to the instances of your tenant. Indeed, a default AZ, which is chosen at random, is now linked with your tenant ID. The default placement is set by default during the creation process. However, if you need to specify a specific AZ for application architecture constraints or DRP purposes, you still can set the target AZ in which to create the new instance.
With this new feature we wanted to help our customers who reported that they had difficulties in choosing the right AZ or keeping with the same choice, getting instances scattered across the AZs of a region. The network latency will also stay minimal inside this AZ. Keeping the default placement choice will entail an homogenous placement for your instances inside the same AZ.
Console
This new feature is available in the console in the screen of ECS instance creation process as shown below. If you do not specify an AZ, the default placement is selected by default. This default placement is applied through the selection of the Random button. It is called Random since the first ECS instance of your tenant will be placed in an AZ selected randomly. All the other ECS instances inside this region will be deployed in the same AZ as the first instance fi the Random AZ is set.
|
Flavor |
vCPUs |
Memory (GiB) |
Maximum/Assured Bandwidth (Gbit/s) |
Maximum PPS (10,000) |
|
d3.xlarge.8 |
4 |
32 |
5/2.5 |
50 |
|
d3.2xlarge.8 |
8 |
64 |
10/5 |
100 |
|
d3.4xlarge.8 |
16 |
128 |
20/10 |
160 |
|
d3.6xlarge.8 |
24 |
192 |
25/15 |
220 |
|
d3.8xlarge.8 |
32 |
256 |
30/20 |
280 |
|
d3.12xlarge.8 |
48 |
384 |
40/32 |
400 |
|
d3.14xlarge.10 |
56 |
560 |
No limit/40 |
500 |
Availability
D3 is generally available. It is now launched in eu-west-0 region (Paris).
Benefits
D3 series offers much better compute price performance over D2.
D3 instances has increased performances for the same amount of vCPU and RAM over D2:
- on the network, with higher maximum throughput, guaranteed bandwidth and PPS by factors between 233 % to 367%.
- on the computing, with higher performances measured by factors between 318% and 735%.
Use Cases
D3 flavors are suitable for applications that require large volumes of data to process, high I/O performance, and rapid data switching and processing. The targeted use cases are MPP data warehouse, distributed MapReduce and Hadoop computing, distributed file systems, network file systems, and log/data processing.
Limitations
D3 has a maximum number of 50 TiB of local SAS disks and 560 GiB of RAM.
Console / API and Resources
As all other instance flavors, the new D3 series is available through the console and API.
More information on ECS disk intensive instances is available here.