News: 1669154407

  ARM Give a man a fire and he's warm for a day, but set fire to him and he's warm for the rest of his life (Terry Pratchett, Jingo)

Intel reveals pay-to-play Xeon features with software-defined silicon

(2022/11/22)


Intel's software-defined silicon service will let organizations pay money to enable features that are hardwired into future Xeon server processors such as Intel Software Guard Extensions, signaling a major shift in how users pay for computer chips.

The x86 giant's software-defined silicon (SDSi) service is officially known as "Intel On Demand," and while the company has not made a proper announcement, Twitter users on Tuesday spotted a [1]new page on Intel's website outlining a list of paid Xeon features, a list of server vendors and IT providers who are supporting the initiative, and information about two ways of using the service.

We have good reason to believe Intel On Demand will be introduced with the 4 th -Gen Xeon Scalable CPUs, code-named Sapphire Rapids, which will [2]launch early next year after several delays. For one, the list of paid Xeon features include those [3]that will debut in Sapphire Rapids , and a few members of the press viewed a [4]recent video where executives discuss 4 th -Gen Xeon serving as a "test ground" for SDSi.

[5]

SDSi has been a great source of curiosity and anxiety in the IT world because of how it will charge users money for features already baked into the silicon of processors. What Intel has not made explicitly clear is whether SDSi will be forced onto users buying next-gen Xeons. A charitable reading of the new Intel On Demand page and a reference to the service in May suggest that SDSi will be available as an option.

[6]

[7]

"The introduction of the Intel On Demand service offering helps enterprises meet the needs of evolving workloads, product sustainability and opportunities to scale systems near the data," Intel said in a [8]May press release , before Intel [9]confirmed in November that Intel On Demand is SDSi's official brand name.

Intel did not immediately respond to a request for comment.

How Intel On Demand will work

According to the new Intel On Demand page, SDSi will be available in two ways.

The first is an activation model, which allows users to enable features on Xeon processors through a one-time activation. This involves a six-step configuration process, where the customer makes an upgrade request to an infrastructure provider, which then shares the request with Intel to receive a license for the customer, according to an official diagram.

[10]

The provider then enables the license in the customer's servers. State information of the customer's SDSi-enabled datacenter is shared with the provider and Intel after.

The Intel On Demand features users can pay to activate consist of:

Intel Software Guard Extensions, a security feature that is meant to protect data in encryption portions of the CPU's memory.

Intel Quick Assist Technology, which can accelerate data encryption and compression workloads by offloading through an offload mechanism.

Intel Dynamic Load Balancer, a hardware-managed load balancing system in the processor that is designed for telecom applications.

Intel In-Memory Analytics Accelerator, which is meant to speed up compression and decompression for big data applications and in-memory analytic databases.

Intel Data Streaming Accelerator, which aims to eliminate bottlenecks in data movements between the CPU cores, memory, caches, attached storage, and networked storage devices.

The latter three features will debut in the upcoming Sapphire Rapids chips.

The second way to use SDSi is a consumption model, which allows users to use Intel On Demand through as-a-service offerings like Hewlett Packard Enterprise's [11]GreenLake platform . This will let organizations "harness Intel Xeon processors to scale capacity, optimize value and dynamically align your infrastructure to meet business demand," according to the Intel On Demand page.

Who's supporting Intel On Demand, and what comes next

Intel lists seven server vendors and IT service providers who will support SDSi's activation model or consumption model, or both. This includes Lenovo, which will use Intel On Demand for its [12]TruScale IT-as-a-service portfolio . Other firms backing the SDSi service include Supermicro, Inspur, H3C, phoenixNAP, and Variscale. Intel's server business, the Datacenter Solutions Group, will also support SDSi.

It's not a surprise to see Intel aligned with server vendors who are pushing as-a-service consumption models, which purport to provide organizations with greater flexibility in how their infrastructure is configured and give them a way to spread out equipment costs, rather than paying for all the kit upfront.

[13]Head of Intel Foundry Services resigns just as chip biz gets going

[14]Chinese Loongson chips coming in 2023, on par with 2020 x86 kit

[15]TSMC confirms 3nm fab in Arizona to make neighbor Intel sweat a little more

[16]Intel hit with $948.8 million VLSI infringement verdict

Intel CEO Pat Gelsinger [17]has previously stated that he wants to significantly increase the company's revenue streams from software and services.

There are still plenty of questions unanswered about Intel's SDSi service, including how the costs will compare between regular Xeon-powered infrastructure and datacenters running SDSi. Besides the reference in a May press release, Intel has been quiet about Intel On Demand from a PR perspective. The only other details have come out through a [18]Linux kernel mailing list and a [19]GitHub page .

However, we do feel certain that SDSi is coming soon, and, unless a major change happens, it will be around for the indefinite future.

[20]

In an October video reportedly seen by [21]Club386 , two senior Intel employees confirmed that the company will introduce SDSi with the upcoming Sapphire Rapids chips.

Intel fellow Ronak Singhal said in the presentation that SDSi will essentially act as a "try-before-you-buy program," letting users test new features before implementing them into applications for production.

Intel datacenter executive Lisa Spelman, in the same video, added that the company plans to "ramp SDSi heavier" in the next generation of Xeon Scalable CPUs. That's likely a reference to Emerald Rapids, which is expected to launch in the second half of 2023, [22]last time we checked . ®

Get our [23]Tech Resources



[1] https://www.intel.com/content/www/us/en/products/docs/ondemand/overview.html

[2] https://www.theregister.com/2022/11/01/intel_sapphire_rapids_q1_2023/

[3] https://www.theregister.com/2021/08/19/intel_architecture_day/

[4] https://www.club386.com/intel-talks-up-oft-delayed-4th-generation-xeon-sapphire-rapids/

[5] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/systems&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2Y31UhVCXhPP0NRAzmzohVgAAABU&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0

[6] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/systems&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44Y31UhVCXhPP0NRAzmzohVgAAABU&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0

[7] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/systems&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33Y31UhVCXhPP0NRAzmzohVgAAABU&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0

[8] https://www.intel.com/content/www/us/en/newsroom/news/vision-2022-news-overview.html#gs.iwjaj6

[9] https://lore.kernel.org/lkml/20221101191023.4150315-1-david.e.box@linux.intel.com/

[10] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/systems&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44Y31UhVCXhPP0NRAzmzohVgAAABU&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0

[11] https://www.theregister.com/2022/06/28/hpe_greenlake_private_cloud/

[12] https://www.theregister.com/2021/09/08/lenovo_throws_everything_its_got/

[13] https://www.theregister.com/2022/11/21/intel_foundry_services_boss_quits/

[14] https://www.theregister.com/2022/11/21/chinese_loongson_chips/

[15] https://www.theregister.com/2022/11/21/tsmc_3nm_fab_arizona/

[16] https://www.theregister.com/2022/11/16/intel_hit_with_9488_million/

[17] https://www.theregister.com/2022/03/11/intel_software_expansion/

[18] https://www.theregister.com/2021/10/19/intel_sdsi/

[19] https://github.com/intel/intel-sdsi

[20] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/systems&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33Y31UhVCXhPP0NRAzmzohVgAAABU&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0

[21] https://www.club386.com/intel-talks-up-oft-delayed-4th-generation-xeon-sapphire-rapids/

[22] https://www.theregister.com/2022/06/08/intel_sapphire_rapids_delay/

[23] https://whitepapers.theregister.com/



devin3782

Thanks Intel, I hate it, there should be law against this sort of crap I find it abhorrent to withhold functionality of a physical object with a software lock (see also Tesla and BMW with heated seats) a pox on their house. This is also thoroughly environmentally unfriendly.

What could go wrong?

redpawn

Rates will of course remain competitive and if Musk or his ilk buys Intel, rest assured your servers will work even better. Genius.

How? Why?

Anonymous Coward

Does this mean the system with an SDSi CPU has to have a service running to communicate with Intel HQ and enable the chip on each reboot? Sounds needlessly complex, never mind the security risks. If problems crop up, especially vulns, I could see this costing more than just hard wiring chips to be forever enabled or not (and why disable when you've got AMD on your heels?).

Subscriptions for broswer soft - I can see that because the browser is sandboxed. The CPU is not.

AMD, ARM or RiscV-poppers to the rescue?

b0llchit

Maybe, just maybe, other vendors might have an edge if they do not go this way. It would be a very nice ad for them... "Our products just work when you pay once . The competition wants you to pay many times."

Risky

Richard 12

It's going to affect their yields, as it means selling chips where the features have been verified as working, but Intel don't know whether the customer will ever actually pay for the feature or not.

I predict that 99.9% of customers either enable it at initial purchase or never, thus effectively turning it into an expensive way of having multiple SKUs.

You'll feel much better once you've given up hope.