ABCDEFGHIJKLMNOPQRSTUVWXYZAA
1
IC MembersOCP
2
NamesDJJessica GJeff CJustin SteadmanJohn LeungLihua YuanSteve MillsJia NingMatt ShumwayCraig WhiteRon MinnichElaine PalmerBill CarterSteve HelvieRajeev SharmaArchna HaylockDirk Van Slyke
3
RolesCo-ChairCo-ChairCo-ChairData Center FacilityHW ManagementNetworkingR&PServerStorageTelcoOSFSecurity
4
Current IntiativesDescription
5
ModularityDeliver scalable, modular components, and open interfaces to enable vanity free HW developmentAligned with the ongoing modularity efforts.

Would like to see BOD companies guide further the direction of modularity based on the needs that can be common across them.		Aligned with the ongoing modularity effortsAgreeI agree with the message. Would it be better worded to say, "Deliver modular components that are scalable with open interaces that enable vanity free HW development."N/AagreeStrongly agree. There is uncertainty around how htis impact storage. So this not impact the storage community. Specifications of modular components and open interfaces are an important aspect to control costs for different form factors; an example includes the OCP NIC/Mezzanine specification.Is an NDA-free SIG process conceivable?I agree with Justin, and would add security. "Deliver modular components that are scalable with open interfaces that enable vanity free HW development. Components incorporate security within the design."Looking back at open interfaces: CXL, Gen-Z, CCIS, CAPI, etc. Each drove their own SIG to define these interfaces. After a few years, the SIG asks OCP about collaboration and help with adoption but was unwilling to partner up front or establish the SIG under the OCP framework (openCAPI and CXL are two examples). There is a much better opportunity to build an enabling plan when we are involved from day 1. I believe OCP can do a MUCH better job of providing the organizational framework for any of these SIG's but we need a formal partnership up front. I believe the strategy is to become the organization of choice for hosting 'open interfaces' and shift away from SIGs. If we can become the best option rather than a separate SIG, the larger OCP community of suppliers, ODMs, end-users can move much faster.
6
At-Scale Operations
Provide common, secure manageability across platforms at cloud scale. Deliver solutions that can be leveraged from the cloud to the edge.This is a compelling application for the large enterprises which may want to use OCP gear to run their own private on-prem cloud. Would be great to have some large enterprise provide thought leadership here.

This can also promote OCP adoption outside board companies.		Aligned with Common ProfilesAgree** Are we limiting our audience with the statement "from the cloud to the edge"? Does this remove an audience such as single tenant or unique environments that may not be "cloud", but could find value in what we offer?
** I do not belive this iniative is properly defined by our description. Both common manageability, and secure manageability are great intiatives, I am just not sure they are listed with the correct intiative.
** Would a better description be something like, "Deliver solutions that enable rapid response to increased demand while maintaining effeciency."		N/ASo how does this get messaged to/integrated with the needs of "enterprise" users like Target / Solution Providers?Strongly agree.

Input: May need to increase stakeholder engagement and reducing the bifurcation of the operation requirements 		This is very important, "at scale" is sufficient as opposed to "at cloud scale". This is probably the single most important aspect of driving costs out of the system. Today opex is often high because of the need for a dedicated staff per HW platform. In telco in particular, it is common for a HW platform management system to become part of the service description, which leads to a subtle form of "lock in". A change in Vendor and/or a change in the service can become very expensive. need to become aware that management (deep knowledge) can have negative imipact on confidentiality. E.g., full jtag access for management could be a very bad idea as it implies "unlimited root." Manageability must not become an attack vector.Agree with Craigat scale' should include large Enterprise size of scale.
7
SustainabilityIncreased standardization on metrics for utilization, supply side resource consumption, and OCP-certified sustainable design practices and operating modesThis is identified as 1 of the 5 strategic initiatives for 2021 (see OCP blog announcement https://www.opencompute.org/blog/ocp-incubation-committee-strategic-plan-for-2021.) Work has started on this front with 2 workstreams identified: LCA and Life cycle extension.

Circularity is also included in this effort to achieve sustainability goals.		This is an important effort that is in its initial scoping phase as a new focus area/technology roadmapSustainability is important should we also look at the inevitable end of life of a product and how easily a design can be recycled, limiting the amount of hazardous material used in a design, etc. ?The goal and description are good, but is the initiative better labeled with Standardization vs. sustainability?N/AThis is aligned with our current efforts in this area, but it would be good to understand if our current scope aligns with the Boards expectationsI am interested to know thoughts around identifying potential sustainability opportunities in creating economic return. Hope this may speed up the progress. Storage is fully behind this initiative and wants to address several key issues. Currently storage devices are crushed on failure making it impossible to effectively recycle and a loss of circularity. Need to define security that enable secure erase of data for device reuse. Also, need to drives aoption of T10/T13 repurpose and depopulate to reduce drive swaps and increased HDD luseable lifespna. Let's not lose sight of how deep this goes, e.g., one time fuses on chips can kill sustainability. Another reason that OSF is critical for this area. I would strike the term "OCP-certified" because we are not a certifying body. The description stands on its own without that term.
Requires hyperscale products 'designed for sustainability' and fullfilled by a managed supply chain to be made available to others. Not the case today. Need for tools and education on how to perform LCAs and publish EPDs.
8
SolutionsImprove scale adoption of OCP designs through streamlined certification of vertical solution stacksOCP is working to incubate Test and Validation efforts that can provide a venue for certification efforts. OCP has previous embarked on certification and badging efforts which did not frutify. Would be good to know what were the issues so we can circumvent them this time around.

OCP also needs to enable third party companies that can provide support services for support and operations of OCP gear for enterprises.		Part of this is Platform Security. Interested to learn more about what else is included in the "solutions" category from the board perspective. This is a very important initiative to help adoption. Certification of a solution is one step with system integration/support services being the second. We should look to engage with partners/members that can perform the integration servicesGreat initiative. The description is unclear with "scale" and "vertical solution stacks". I am not saying they should be removed, just challenging the need for them. It is very possible I am just unaware of their meaning but I feel they make the message confusing and our description would be just as clear without them. N/ASuggest to involve potential adopters / users in this processSuggest to involve potential adopters / users in this processThis really needs clarity on the benefits of providing an OCP solutiuon. Use cases on how providing an OCP solution has increased business. The were some questions around the repoirt in 2020 about the revenue of OCP products. Need ot clear examples of solutions being adopted across multiple companies I would not call this a "streamlined certification". As mentioned in another response, "component certification" has proved to be difficult. However, a "reference implementations" may be a lower bar to make progress. I think this is also tied into manageability. There are "service" vertical systems ("a stack") but an inclusive stack that covers discovery, provisioning, in-life maintenance, and deprovisioning is more challenging. Also OCP needs a robust software control system for this effort.As we can see with OSF, this will likely require handholding. I think some deeply technical "design days" at the summit where we cover important areas, and how to meet a checklist, might help. Also, some sort of ambassador role from OCP to the companies -- much deeper engagement than supplier survey -- might help. The ODMs are under market pressure to deliver first and that can weigh against the OCP goals (again, as we have seen with OSF). It would help for the Summit to move slightly away from one-way talks to almost a hackerthon mentality -- bring your soldering irons!Ditto Craig. Again, we need to avoid the term "certification". Open source reference implementations are always welcome and speed adoption, if done properly with unrestricted licenses. For full stacks, however, lining up all the licenses would be a legal nightmare.Ideally, OCP products are easier to adopt and deploy becuase some level of system testing has been performed and reported, and continues as the stack changes with time. Avoid the term 'certification'
9
10
11
Future InnovationDescription
12
CoolingDeliver best in class, sustainable immersion cooling designs from cloud to edgeOCP ACS project streams are addressing cold plate and immersion approaches.Aligned with this area. Agree. The community has been focusing on the first step: Cloud. More work is needed for Cloud as well, but going forward closer harmonization is needed with Edge IT equipment and Edge Facility Deployments. I have brought this up in the immersion sub-project as future topic. Not sure who and where Edge facility Deployment/Development falls?!Agree we should remove immersion and not limit ourselvesWe should remove "immersion". I do not feel we need to limit this topic. I also question the "from cloud to edge" and ask if we are limiting our audience and/or is there a need to add this to our statement?N/AWe are currently supporting projects working on both cold plate and immersion specifications. I think there are still long term applications at scale for both of these technologies. Is there a reason not to support liquid cooled solutions?I suggest to not limit this work to immersion cooling which is one of the many emerging cooling solutions Will imersion cooling be adopted for storage solutions? What is the value proposition. The case fro compute is well understood, but happens to storage in this environment. Cooling of dense platforms, at scale while simultaneously driving/keeping costs down is important.no commentNo commentDon't limit to immersion
13
OpticsLead in defining market requirements for process and technology transitions for optimal convergenceOCP will need to find way to work with OIF to create a momentum here. OIF has all the suppliers on optics side who need to be present within OCP community to drive the marketplace forward.

Optics is also a very broad space, so we need better understanding of what BOD is thinking when they refer to Optics. Is this developing modules, lasers and optical components, optical gear etc.		It would be good to have OCP play a larger role here. How does this align with other already existing industry forums? Do we need to create partnerships with those forums?There are the optical components themselves and then how these are used in a designs ( plugin modules, OBO, CPO, etc.) which (or all?) of these is the initiative focused on? N/AThis is definitely of interest to us, but I am not sure we are actively working on this that I am aware.I would like to learn more about what role OCP(foundation and community) want to play in the OCP 2.0 framework with other industry forums. E.g. leading ? supporting? specification definition? requirement definition?
Optical HBAs for storage systems.? So OCP's role here is showing there is a market for our requirements or ... not sure I see OCP driving improvements in glass :-)?No commentWith hyperscale driving 80% of the demand, the question i have is what information/demand forecast can hyperscalers' share with each other and the industry to actually guide resource deployment. i like the desc but recognize the action come from the board companies.
14
ChipletsHelp define interfaces for future co-packaging to enable best-in-class components from the silicon up -- COMBINE WITH OPEN SILICON INITIATIVE. TOO CLOSELY CONNECTEDOCP ODSA is covering this space and momentum is building here. The key for ODSA to succeed is to deliver an open die-to-die interface for the community to develop chiplet marketplace quickly. For this we are working to have a BOW test chip defined. We need some amount of funding commitments from the value chain stakeholders (from builders to users) to ensure the companies have "skin in the game" for developing something that is useful.Is the goal to generate "modular" designs using chiplet technology?AgreeN/AThis is an area that can provide much benefit in future years. Maybe OCP needs to create a central framework to enable this since the actual defnitions/specifications will occur in different project workgroups (Network, server, storage, telco)Fits well with OCP 2.0 framework as a long term effort - seeing good momentums in definition and proof-of-concept.

Needs to find stream of value in productionlized work, or steps towards that in the next a few years. 		No commentI've learned that all vendors view interconnect at this level as crown jewel IP. it's not like a motherboard. I see this as a real challenge -- standardizing on something they view as a differentiator. Fingers crossed.No comment
Need a joint funding model to validate such interfaces. Suggest combining chiplets and open silicon. Chiplet development becomes a interim step to full silicon development.
15
Open SiliconDrive tools and reference platform standards for faster innovation in open silicon interfaces - USE AI/ML FUTURE INITIATIVE WORKSTREAM as an example.Unclear on this topic. Tools refer to EDA tools, simulation tools, IC design tools or something else. I feel this is a stretch beyond OCP community capabilities and not sure what we are accomplishing by including SW tools into OCP ecosystem. Further clarification is needed for OCP IC to evaluate.Not sure what this means. Clarification requested.The focus of this is not clear to me N/AHow does this differ from the chiplets? Same idea but at a larger scale?Is it about EDA tools and simulation/emulation platforms? Not sure if this is the most urgent pain point. Is storage compute an architecture that can scale? Need submissions around that activitiy or statements of what architecture would work. This one is also really hard. Here's a simple one -- all general purpose IO comes with patent encumbrement. So it's hard to design an open IP to read a power button. Does this include DDR5 silicon, for example? What are the boundaries here?I don't like the last word "interfaces". The industry doesn't really have a success story with open silicon yet (maybe RISC5). IMO, if we can bring together companies that share a common problem, we can build silicon to solve it. That silicon solution will be very specific and doesn't need an open interface to be successful. I'm even wondering if the problem can be solved by defining a set of behavioral models and the API.
16
17
General commentsIt is great start to have a dialogue and a forum for OCP IC and OCP BOD to interact. This allows for better coordination and growth of the OCP community which is aligned with broad goals derived from OCP BOD. We thank the BOD for being pro-active in sharing their thoughts and look forward to continued communication to grow OCP in a deliberate and measured manner that is both impactful and successful.I really appreciated hearing from the OCP board of their vision and looking forward to closer alignment between board and IC. We do need to map out what is different in the board's vision to what we are currently doing to ensure that we are addressing those additions.Great to have the communication path open :-)> I am glad there is a dialogue between the Board and the IC and provide the IC visibility. > Is the IC to wait for direction from the Board regarding where it wants IC help? > The IC has created its own list of areas of interest. How do the Board and IC lists mesh?I appreciate OCP BOD reaching out in aligning and confirming direction with ICsIt is important to keep in mind that we are a member driven organization. These members includes component, packaging (oXm's), solution providers, and adopters. Our mission needs to be cognizant of each type of member and to articulate the "value add" to each member. At the end of the day the OCP providing value to these members is what provides the mental horsepower to drive our mission. I worry about trying to do everything at once. This is challenging. Thank you for the opportunity to provide feedback. One item I don't see is supply chain integrity. I use integrity rather than security, because security implies too much. For example, integrity could incorporate standards for bills of materials, which need to be protected from tampering, but the end goal is not security. The end goal is to get the information from the supplier to the consumer in a format that the consumer can process and trust.
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100