Intel Data Center Blocks - ordered, received, built
Experienced administrators are well aware of Intel server hardware. For many years, the company has been manufacturing both ready-made server platforms and various components for them. This topic rarely appears in our blog, giving way to more media, but now we would like to pay attention to it and tell about the Intel Data Center Blocks (Intel DCB) program and the new, just released equipment that it covers.
The thesis that the principles of building server systems have changed significantly over the past 10 years has already become a commonplace. In light of the increased loads and their unification, density, scaling and unification came first. In other words, from individual servers that perform certain tasks, we moved to the "bricks" to build a homogeneous, software-defined system. In the wake of the general trends in 2016, Intel launched Intel Data Center Blocks , which streamlined the server line in accordance with this concept.
A few months ago, a new generation of Intel DCB platforms was released. Let's use their example to look at the current state of affairs in the line. To date, the system node can be made in two versions:
Server family R1000WF
Let's take a look at the characteristics of the site as an example - they are typical for a sample server of 2017.
Similarly, the characteristics of the processing unit for the chassis. The chassis allows the modules to share the disk subsystem and power bus. In addition to platforms, Intel also released a new generation of RAID and HBA controllers with various combinations of internal and external ports.
The set of proprietary technologies, given in the last row of the table (and in fact there are much more), is one of the parts of the Intel Data Center Blocks strategy; they provide good behavior and maximum return node in the general system of "bricks". However, the question arises: in one form or another, most of this functionality has been implemented by other manufacturers; in addition, the server configuration, as already mentioned, is also typical for modern models. What is so special about Intel?
In addition to the technological aspect, the Intel DCB program includes a manufacturer-customer interaction system Configure To Order (CTO) Tool. Unlike conventional service systems, the end product of which is the position list (BOM), sent to the Intel DCB CTO reseller, you can directly operate with manufacturer quotas that are sold to the customer through an authorized sales representative.
Intel DCB CTO Tool Interface
Thus, the whole complex of equipment is supplied, one can say, on a turn-key basis, according to manufacturer quotas. This avoids the situation “this came, and this is not” when you rely on the distributor. Reduced delivery and deployment time, which is also money. So, as they say, you will pass by - pay attention to Intel Data Center Blocks. It seems to be worth it.
The thesis that the principles of building server systems have changed significantly over the past 10 years has already become a commonplace. In light of the increased loads and their unification, density, scaling and unification came first. In other words, from individual servers that perform certain tasks, we moved to the "bricks" to build a homogeneous, software-defined system. In the wake of the general trends in 2016, Intel launched Intel Data Center Blocks , which streamlined the server line in accordance with this concept.
A few months ago, a new generation of Intel DCB platforms was released. Let's use their example to look at the current state of affairs in the line. To date, the system node can be made in two versions:
- The classic two-socket server with a height of 1U (earlier this year, single-socket servers of the initial level were also released);
- 2U chassis, accommodating up to 4 Intel Compute Module modules - dual-socket microservers.
Server family R1000WF
Let's take a look at the characteristics of the site as an example - they are typical for a sample server of 2017.
| CPU | 2 x Intel Xeon Scalable |
| Chipset | Intel C624 |
| Nutrition | 2 x 1100 W with redundancy |
| Network interfaces | 2 x 10GbE LAN |
| Storage | 8 x 2.5 ”outside, 2 x M.2 inside |
| PCIe configuration | 2 x16, 2 x24 (optional) |
| Memory | 24 x DDR4-2666 MHz |
| Intel technology support | Intel Optane Memory
Intel Node Manager Intel On-Demand Redundant Power Intel Advanced Management Technology Intel Server Customization Technology Intel Build Assurance Technology |
The set of proprietary technologies, given in the last row of the table (and in fact there are much more), is one of the parts of the Intel Data Center Blocks strategy; they provide good behavior and maximum return node in the general system of "bricks". However, the question arises: in one form or another, most of this functionality has been implemented by other manufacturers; in addition, the server configuration, as already mentioned, is also typical for modern models. What is so special about Intel?
In addition to the technological aspect, the Intel DCB program includes a manufacturer-customer interaction system Configure To Order (CTO) Tool. Unlike conventional service systems, the end product of which is the position list (BOM), sent to the Intel DCB CTO reseller, you can directly operate with manufacturer quotas that are sold to the customer through an authorized sales representative.
Intel DCB CTO Tool Interface
Thus, the whole complex of equipment is supplied, one can say, on a turn-key basis, according to manufacturer quotas. This avoids the situation “this came, and this is not” when you rely on the distributor. Reduced delivery and deployment time, which is also money. So, as they say, you will pass by - pay attention to Intel Data Center Blocks. It seems to be worth it.
All Articles