OCP NIC 3.0 Spec update ready for review - TSFF and editorial changes - version 1.1.1


McGraw, Keegan
 

Hello NIC Community,

 

As a new contributor to the NIC subgroup, a couple of comments noted on the V1.1.1 spec we would like to get group feedback:

 

  • The below table shows specifically puts QSFP-DD as a TSFF solution only. Mechanically, I believe this is not true as it is the same height as QSFP. For Passive DAC Switch/NIC connections, QSFP-DD is still viable in SFF. Within the thermal work group we are still investigating the viability of QSFP-DD in SFF. While the results don’t look as encouraging, we still haven’t concluded that QSFP-DD will not work in SFF. From a mechanical point of view, I do not see why SFF cannot accommodate QSFP-DD, especially as AOC power comes down over generations and thermal solutions improve. – Page 34
    • We propose to include QSFP-DD as a SFF, TSFF option while relying on the Thermal Simulation and Airflow requirements at the system level to determine viability.

 

 

  • Based on the recent thermal simulations the thermal workgroup has been performing, the airflow for hot aisle applications ranges from 500-1000 LFM. However; the table in the spec shows that this is beyond the high end limit (300 LFM). Might this might be an opportunity to adjust these limits? – page 193
    • Need some system vendor input who have Hot Aisle Applications

 

 

Thank You,

 

Keegan McGraw

Technologist

(317)-496-7070

 

From: OCP-NIC@OCP-All.groups.io <OCP-NIC@OCP-All.groups.io> On Behalf Of Damien Weng Kong Chong via groups.io
Sent: Tuesday, July 20, 2021 12:04 AM
To: Ng, Thomas1 <thomas1.ng@...>; OCP-NIC@OCP-All.groups.io; OCP-Server@OCP-All.groups.io
Cc: Ng, Thomas1 <thomas1.ng@...>; damien.chong@...
Subject: [OCP-Mezz] FW: OCP NIC 3.0 Spec update ready for review - TSFF and editorial changes - version 1.1.1
Importance: High

 

Sent by an external sender


Hi OCP NIC and OCP Server members,

 

We are finally ready to open up Tall Small Form Factor (TSFF) spec review.

We are doing NIC spec contributors and NIC/server member review concurrently for this change as we believe contributors are well aligned after months of discussions.

We request your help to provide feedback to Thomas and myself by Wednesday August 4th, 2021.

 

OCP NIC wiki: https://www.opencompute.org/wiki/Server/Mezz

Direct link to 1v11 release candidate with change bar: http://files.opencompute.org/oc/public.php?service=files&t=0e91c3edf08e352e03923066f8474f74

Direct link to 1v11 release candidate without change bar: http://files.opencompute.org/oc/public.php?service=files&t=4767d9e7cf6c3ea86e82ae34c1ac10fe

 

For high level changes, please refer to email from Thomas below.

 

Thanks in advance!

 

 

 

Damien Chong

Facebook | Infrastructure | Technical Lead Manager | Email: dchong@... | Tel: +1 650 250 6582

 

--When I am in cc list, the email would be read at lower priority-- 

 

From: Ng, Thomas1 <thomas1.ng@...>
Sent: Monday, July 19, 2021 7:41 PM
 

 

Good afternoon all OCP NIC 3.0 Workgroup Participants –

 

The OCP NIC 3.0 spec has been updated with the TSFF information as well as editorial changes. DRAFT version 1.1.1 is ready for your review.  Per the review guidelines set with release 1.0, the document is open for a two week review period. Please e-mail me with your comments – this can be a list, or in-line comments in the Word source. Your review is requested by Wednesday August 4th, 2021.

 

You can find the document source, and corresponding change bar / clean versions of the PDF on the OwnCloud site.

 

URL

http://files.opencompute.org/oc/public.php?service=files&t=ea72c1c3050057655da8eda1f42dd0e2&path=//OCP%20NIC%203.0%20-%20Spec%20%28docx%2C%20pdf%29

Word Doc (Source)

OCP_NIC_3.0_d1v11_20210719c_TN.docx

PDF (change bars)

OCP_NIC_3.0_d1v11_20210719c_TN_CB_DRAFT.pdf

PDF (clean)

OCP_NIC_3.0_d1v11_20210719c_TN_no_CB_DRAFT.pdf

 

The high level list of changes are shown below. Please refer to the specific sections for all change details:

 

- General – Changed “SFF” references to “SFF/TSFF” where applicable.

- General – Format all tables with repeated headers when they span multiple pages

- Section 1.2 – Added Molex to the acknowledgements section.

- Section 1.4 – Added TSFF, OSFP-RHS and QSFP-DD to the acronyms section.

- Section 1.6, 2.1.1 – Added text for TSFF and clarify TSFF shares the same mechanical outline as SFF.

- Section 1.7.1 – Added TSFF to the OCP NIC 3.0 Form Factor Dimensions table; add Max Z-height information.

- Section 1.7.2.x - Updated text "Up to PCIe Gen 5 (32 GT/s) support"

- Section 2.2 - Add QSFP-DD, OSFP-RHS for TSFF in Line side implementation. Add QSFP-DD, OSFP to standards cross reference table.

- Section 2.4.1, 2.4.2 – Updated faceplate subassembly drawing and BOM callouts.

- Section 2.4.4 – Added TSFF generic faceplate drawings. Drawings only contain dimensional deltas from SFF.

- Section 2.5.2 – Added TSFF keep out zones. Drawings only contain dimensional deltas from SFF.
- Section 2.8.5 - Replaced Figure 56 - SFF Baseboard Chassis CTF dimensions (Rear Rail Guide View) as it was clipped in the previous release.

- Section 2.8.6, 2.8.7, 2.8.8, 2.8.9 – Added TSFF CTF dimensions for pull tab, ejector and internal lock board dimensions. Add TSFF baseboard CTF dimensions. Drawings only contain dimensional deltas from SFF.
- Section 2.10 - Added TSFF Generic faceplate assembly CAD.

- Section 3.4.2, 3.5.4 – Clarified that the BIF[2:0]# pins shall not change states after AUX_PWR_EN is asserted.

- Section 3.4.4 – Clarified that the RBT_ISOLATE# signal shall not change state even when the NIC_PWR_GOOD signal is in the don't care state.

- Section 3.7.3 – Added note regarding the use of light pipes above I/O cage assemblies for TSFF and cage implementations.

- Section 3.11 – Clarified power down sequencing diagram - +12V_EDGE may optionally be disabled in ID Mode. This mirrors the same behavior for the power up sequencing diagram.

- Section 5.1 – Added note stating the SFF and TSFF timing parameters are the same since the cards may share the same PBA.

- Section 6.2.2, 6.2.5 – Added placeholder headings for the hot and cold aisle cooling graphs for TSFF.

- Section 6.4.2 – Added placeholder heading for the TSFF thermal test fixture.

- Section 6.4.5 – Added placeholder figure headings for the TSFF candlestick air velocity vs volume flow graphs.

- Section 7.1.3 – Updated Safety requirements. 60950-1 was withdrawn as of 12/20/2020. 62368-1 replaced 60950-1.

 

We’ll roll this document to version 1.2 when the review is completed.

 

Thank you,

Thomas

 

Thomas Ng
Hardware Engineer | Ethernet Networking Division

Intel Corporation | Mailstop JF3-414 | 2111 Northeast 25th Avenue | Hillsboro, OR, 97124

e thomas1.ng@... | d 503 712 2905

 

 






CONFIDENTIALITY NOTICE: This message (including any attachments) may contain Molex confidential information, protected by law. If this message is confidential, forwarding it to individuals, other than those with a need to know, without the permission of the sender, is prohibited.

This message is also intended for a specific individual. If you are not the intended recipient, you should delete this message and are hereby notified that any disclosure, copying, or distribution of this message or taking of any action based upon it, is strictly prohibited.

English | Chinese | Japanese
www.molex.com/confidentiality.html


Ng, Thomas1
 

Good morning Keegan –

 

Thank you for your input. I’ll defer to the mechanical and thermal group for input on the cage vs SFF/TSFF implementations as well as the airflow.

 

I’ve fixed the cross to the LFF Lineside implementation in the local draft. It should read “Section 2.4.5 for LFF”

 

-Thomas

 

Thomas Ng
Hardware Engineer | Ethernet Networking Division

Intel Corporation | Mailstop JF3-414 | 2111 Northeast 25th Avenue | Hillsboro, OR, 97124

e thomas1.ng@... | d 503 712 2905

 

 

From: OCP-NIC@OCP-All.groups.io <OCP-NIC@OCP-All.groups.io> On Behalf Of McGraw, Keegan
Sent: August 3, 2021 12:32 PM
To: OCP-NIC@OCP-All.groups.io; Ng, Thomas1 <thomas1.ng@...>; OCP-Server@OCP-All.groups.io
Cc: damien.chong@...
Subject: Re: [OCP-Mezz] OCP NIC 3.0 Spec update ready for review - TSFF and editorial changes - version 1.1.1

 

Hello NIC Community,

 

As a new contributor to the NIC subgroup, a couple of comments noted on the V1.1.1 spec we would like to get group feedback:

 

  • The below table shows specifically puts QSFP-DD as a TSFF solution only. Mechanically, I believe this is not true as it is the same height as QSFP. For Passive DAC Switch/NIC connections, QSFP-DD is still viable in SFF. Within the thermal work group we are still investigating the viability of QSFP-DD in SFF. While the results don’t look as encouraging, we still haven’t concluded that QSFP-DD will not work in SFF. From a mechanical point of view, I do not see why SFF cannot accommodate QSFP-DD, especially as AOC power comes down over generations and thermal solutions improve. – Page 34
    • We propose to include QSFP-DD as a SFF, TSFF option while relying on the Thermal Simulation and Airflow requirements at the system level to determine viability.

 

 

  • Based on the recent thermal simulations the thermal workgroup has been performing, the airflow for hot aisle applications ranges from 500-1000 LFM. However; the table in the spec shows that this is beyond the high end limit (300 LFM). Might this might be an opportunity to adjust these limits? – page 193
    • Need some system vendor input who have Hot Aisle Applications

 

 

Thank You,

 

Keegan McGraw

Technologist

(317)-496-7070

 

From: OCP-NIC@OCP-All.groups.io <OCP-NIC@OCP-All.groups.io> On Behalf Of Damien Weng Kong Chong via groups.io
Sent: Tuesday, July 20, 2021 12:04 AM
To: Ng, Thomas1 <thomas1.ng@...>; OCP-NIC@OCP-All.groups.io; OCP-Server@OCP-All.groups.io
Cc: Ng, Thomas1 <thomas1.ng@...>; damien.chong@...
Subject: [OCP-Mezz] FW: OCP NIC 3.0 Spec update ready for review - TSFF and editorial changes - version 1.1.1
Importance: High

 

Sent by an external sender


Hi OCP NIC and OCP Server members,

 

We are finally ready to open up Tall Small Form Factor (TSFF) spec review.

We are doing NIC spec contributors and NIC/server member review concurrently for this change as we believe contributors are well aligned after months of discussions.

We request your help to provide feedback to Thomas and myself by Wednesday August 4th, 2021.

 

OCP NIC wiki: https://www.opencompute.org/wiki/Server/Mezz

Direct link to 1v11 release candidate with change bar: http://files.opencompute.org/oc/public.php?service=files&t=0e91c3edf08e352e03923066f8474f74

Direct link to 1v11 release candidate without change bar: http://files.opencompute.org/oc/public.php?service=files&t=4767d9e7cf6c3ea86e82ae34c1ac10fe

 

For high level changes, please refer to email from Thomas below.

 

Thanks in advance!

 

 

 

Damien Chong

Facebook | Infrastructure | Technical Lead Manager | Email: dchong@... | Tel: +1 650 250 6582

 

--When I am in cc list, the email would be read at lower priority-- 

 

From: Ng, Thomas1 <thomas1.ng@...>
Sent: Monday, July 19, 2021 7:41 PM
 

 

Good afternoon all OCP NIC 3.0 Workgroup Participants –

 

The OCP NIC 3.0 spec has been updated with the TSFF information as well as editorial changes. DRAFT version 1.1.1 is ready for your review.  Per the review guidelines set with release 1.0, the document is open for a two week review period. Please e-mail me with your comments – this can be a list, or in-line comments in the Word source. Your review is requested by Wednesday August 4th, 2021.

 

You can find the document source, and corresponding change bar / clean versions of the PDF on the OwnCloud site.

 

URL

http://files.opencompute.org/oc/public.php?service=files&t=ea72c1c3050057655da8eda1f42dd0e2&path=//OCP%20NIC%203.0%20-%20Spec%20%28docx%2C%20pdf%29

Word Doc (Source)

OCP_NIC_3.0_d1v11_20210719c_TN.docx

PDF (change bars)

OCP_NIC_3.0_d1v11_20210719c_TN_CB_DRAFT.pdf

PDF (clean)

OCP_NIC_3.0_d1v11_20210719c_TN_no_CB_DRAFT.pdf

 

The high level list of changes are shown below. Please refer to the specific sections for all change details:

 

- General – Changed “SFF” references to “SFF/TSFF” where applicable.

- General – Format all tables with repeated headers when they span multiple pages

- Section 1.2 – Added Molex to the acknowledgements section.

- Section 1.4 – Added TSFF, OSFP-RHS and QSFP-DD to the acronyms section.

- Section 1.6, 2.1.1 – Added text for TSFF and clarify TSFF shares the same mechanical outline as SFF.

- Section 1.7.1 – Added TSFF to the OCP NIC 3.0 Form Factor Dimensions table; add Max Z-height information.

- Section 1.7.2.x - Updated text "Up to PCIe Gen 5 (32 GT/s) support"

- Section 2.2 - Add QSFP-DD, OSFP-RHS for TSFF in Line side implementation. Add QSFP-DD, OSFP to standards cross reference table.

- Section 2.4.1, 2.4.2 – Updated faceplate subassembly drawing and BOM callouts.

- Section 2.4.4 – Added TSFF generic faceplate drawings. Drawings only contain dimensional deltas from SFF.

- Section 2.5.2 – Added TSFF keep out zones. Drawings only contain dimensional deltas from SFF.
- Section 2.8.5 - Replaced Figure 56 - SFF Baseboard Chassis CTF dimensions (Rear Rail Guide View) as it was clipped in the previous release.

- Section 2.8.6, 2.8.7, 2.8.8, 2.8.9 – Added TSFF CTF dimensions for pull tab, ejector and internal lock board dimensions. Add TSFF baseboard CTF dimensions. Drawings only contain dimensional deltas from SFF.
- Section 2.10 - Added TSFF Generic faceplate assembly CAD.

- Section 3.4.2, 3.5.4 – Clarified that the BIF[2:0]# pins shall not change states after AUX_PWR_EN is asserted.

- Section 3.4.4 – Clarified that the RBT_ISOLATE# signal shall not change state even when the NIC_PWR_GOOD signal is in the don't care state.

- Section 3.7.3 – Added note regarding the use of light pipes above I/O cage assemblies for TSFF and cage implementations.

- Section 3.11 – Clarified power down sequencing diagram - +12V_EDGE may optionally be disabled in ID Mode. This mirrors the same behavior for the power up sequencing diagram.

- Section 5.1 – Added note stating the SFF and TSFF timing parameters are the same since the cards may share the same PBA.

- Section 6.2.2, 6.2.5 – Added placeholder headings for the hot and cold aisle cooling graphs for TSFF.

- Section 6.4.2 – Added placeholder heading for the TSFF thermal test fixture.

- Section 6.4.5 – Added placeholder figure headings for the TSFF candlestick air velocity vs volume flow graphs.

- Section 7.1.3 – Updated Safety requirements. 60950-1 was withdrawn as of 12/20/2020. 62368-1 replaced 60950-1.

 

We’ll roll this document to version 1.2 when the review is completed.

 

Thank you,

Thomas

 

Thomas Ng
Hardware Engineer | Ethernet Networking Division

Intel Corporation | Mailstop JF3-414 | 2111 Northeast 25th Avenue | Hillsboro, OR, 97124

e thomas1.ng@... | d 503 712 2905

 

 

 





CONFIDENTIALITY NOTICE: This message (including any attachments) may contain Molex confidential information, protected by law. If this message is confidential, forwarding it to individuals, other than those with a need to know, without the permission of the sender, is prohibited.

This message is also intended for a specific individual. If you are not the intended recipient, you should delete this message and are hereby notified that any disclosure, copying, or distribution of this message or taking of any action based upon it, is strictly prohibited.

English | Chinese | Japanese
www.molex.com/confidentiality.html