When to take your foot off the gas

It has been a really difficult year. I try not to indulge in self-pity, but I am a fan of self-reflection for the general purpose of attempting to learn from my experiences. Some of what made the last year difficult was out of my control (my nephew’s cancer diagnosis and fight, our wonderful nanny moving on to a family with infants since my boys are definitely not in that category anymore, etc.), but a lot of it was my choice. Sometimes I wonder why I am so intent on pushing when situations are already stressful. I suspect human nature here, but also likely a special flavor of masochism. I self-identify as doing hard things, and sometimes, wow that makes things hard!

I started a new job a year and a month ago. The transition to this type of role was something I had contemplated for a long time–I love semiconductors and feel like we are in one of the most interesting phases of innovation within design and manufacturing that will occur in my lifetime, but having experienced Silicon and system design without ever having had to operate those systems and understand what it takes to resolve the issues that occur at scale was a hole in my development. I knew there were aspects of design tradeoffs and decisions I could not reason about fully from the feedback of my customers and partners alone. 1+ year at Cloudflare, and I am still learning daily from the teams who keep our network up and running how to build better systems and how to help our vendors design more optimal products for services like ours. In 6 weeks I learned more at Cloudflare about the realities of managing systems at scale than I learned in 20 years of development and design. And I learn more with every project everyday. I will never stop loving the fundamentals of design and manufacturing, but I have greater humility and I hope insight from this experience.

My first day of orientation was also the date of my nephew’s first brain surgery to remove a baseball sized tumor from his brain. My first month was punctuated by PICU text messages from my sister. While trying to learn a whole new suite of products, understand our current projects across my various teams, develop relationships with my key partners and vendors, understand internal processes, build my team, and formulate my vision, I also was trying to make sure my nephew was getting the care he needed, and my sister and brother-in-law got some support. I am VERY lucky to have a large family of sisters and aunties as well as amazing parents, and everyone rallied to this cause, but it was a constant weight. Then the house remodel we had been planning for nearly two years finally kicked off (with us camping in the house because the rents in the Bay Area are no joke!) So I added the “fun” of cooking in my outdoor kitchen at least three nights a week during the wettest winter I have ever seen here in California.

Throughout all of this I remained the Chairperson and President of the Open Compute Project Foundation–hiring our new CEO, helping oversee the vision and strategy, and ensuring that we have tools and capabilities in the staff and infrastructure to support a healthy and vibrant community. I also remained the singer and primary booker for my band, Sinister Dexter, and you know just for fun (or honestly stress management) ran 4 half marathons (including a PR post having babies).

Then in September, I had to find a new au pair, our first match did not work out, and the immigration process for the second meant we went with part-time and strung together childcare for 2 months while balancing the schedules of two working parents. Seriously there were days when I thought to myself, how about I call in sick? I didn’t. I took two one-week vacations with my family and one with friends who have become family, and these were precious, did weekend visits to see my nephew and my sister, and I did actually get sick enough one week that I stopped working at 4 pm to take a bath and pass out, but otherwise I didn’t miss a day.

I don’t say this because I am proud of it–there are times when you need to take a step back and take care of yourself, but the insidious thing about cancer is that the process of fighting is a lot of work for the individual and their direct support team (my sister and brother-in-law) and for the rest of us there was a lot of waiting, hoping, praying, and worrying given that we couldn’t be in the hospital due to covid19 precautions (still!) I don’t like to worry. I’d rather do things, and honestly that was why I consciously chose to stay SO busy, but I am not going to lie that there is a lot of internal reflection going on right now. We got the blessed news in February that my nephew is officially in remission, and the precious dinners, and family nights with him over the last few months are particularly sweet. I want to say yes to the trips, visits, TIME together to make memories. I want to have family dinners around my new kitchen table. I want to be with the people I love and take a deep breath, and maybe, just maybe, not run up a hill in the rain for the next few months.

Stay tuned for a slightly chiller Rebecca. There are times when we just have to take a step back in order to be able to step forward. I think this is finally the time.


International Women’s Day

Today is International Women’s Day. You may not realize this, but as of 2022 women held only 28% of roles in tech with even lower representation at higher grade levels. In data, security, and infrastructure, there are even fewer of us.

I got where I am today not only because of the strong women in my life, but very much because of the men who supported me, particularly early in my career and even now when I am not in the room. Advocacy matters, and we cannot change the ratio without the entire tech community stepping up to embrace equity. Equity is not the same as equality–this is a distinction that Cloudflare has particularly educated me upon.

Equality means each individual or group of people is given the same resources or opportunities.

Equity recognizes that each person has different circumstances, and allocates the exact resources and opportunities needed to reach an equal outcome.

In tech we need to do a better job of bringing people up and into the dialogue. I fully thank the bosses, colleagues, and partners in my life who held the door open for me. I see you and I am grateful. AND we need to do more to make sure the doors are easier for our next generation.

Today, and every day, those of us in leadership need to encourage, enable, and support other women and underrepresented minorities to stay engaged in tech despite the challenges of being the “only” or the “few.” We are stronger and better together. This means making sure if you observe bias you call it out, and speak up for equity. It also means recognizing the additional burden agents of change feel and giving them support.

That also means stating my commitment to others: for those of you earlier in your careers who are struggling, know that you are not alone. Please ask for help, reach out, and know I am here to help. Also, you need to help yourself. Build your support system, your sponsors, your mentors, and your advocates, so you can continue to thrive.

In life we will all face bias. You will be underestimated. The world will not always be fair. For the record, you too have bias. I was recently confronted in a meeting for speaking about a leader whose story inspired me, and reminded by a fellow woman in tech that leader was incredibly racist and very much a product of imperialism. Our history is not perfect, and it is hard to have a discussion about any historical figure who would measure up to our current ideals. Do we stop focusing on the lessons worth learning from history because the people are flawed? I don’t believe we can, AND I’m glad that person spoke up to help me see the pain it caused her, so I could be more nuanced in my commentary. Every day we get the chance to learn, if we make a space where people are willing to speak up and share their differences of opinion.

I personally find faith in one simple fact: we learn more in times of struggle than we do in success. We can only make the world better if we stay in the fight. It is a struggle to be different, to be a change agent, and advocates can speak up for you, but they cannot know what it is to walk in your shoes. All we can do is come together, hold each other’s hands, and give each other grace to learn and grow, so we can keep moving forward together.

Book Club

February Book Review: The Light We Carry

For February book club we read The Light We Carry by Michelle Obama. At its heart this book is a call to action and a message of hope in times of uncertainty, and while it isn’t necessarily the prototypical choice for a book club on leadership in technology, I have historically found biographies and autobiographies of leaders managing turbulent times to be incredibly inspiring. This book was no exception. The book isn’t really memoir, but it is autobiographical in the sense of the stories she shares. Here are the nuggets that particularly resonated with me and the group, and might be lessons leaders in technology want to reflect upon as well:

  1. Build your “kitchen table”. Mrs. Obama talks about the kitchen table of her youth: how many people were gathered there with love and support for one another, not just family, but also friends, and that was particularly poignant to me. With Covid19 lockdowns, increased remote work, the age and associated obligations of my children, and other factors, my “kitchen table” is not as full as I personally would like. The de facto friendships I used to get at work are much harder to create remotely and without the regular touchpoints of work travel making it easier to say “yes” to grabbing dinner or drinks with friends. I try to stay in touch with friends when they happen to be in town, or if I happen to be traveling to their towns, but it is far more sporadic, and the friends and colleagues I once spoke with constantly, of course have less frequent touch points with me now that we are dispersed to other companies. Mrs. Obama shares wonderful stories about her “bootcamp” weekends with her best friends (at Camp David no less) and I have to admit that it was a personal inspiration to remember that we can choose to continue our friendships, and it has immense dividends to us personally and professionally. We need a village, and it begins by picking up the phone, texting to check in, or just sending a little virtual love on social media to the folks we care about. The limiter here is time of course, but a book club, email/text thread with besties, running/gym club/buddy, etc. can be good life hacks to keep folks in touch on a semi-regular basis. There is a great book called Multipliers, which changed my life on this topic, and I know lots of other folks refer to Never Eat Alone. There is never enough time to do all the things, but if you combine objectives with friendship, there absolutely can and will be a multiplier effect, and increased accountability that results in far greater dividends throughout life.
  2. Gladness is nourishing. Mrs. Obama refers to a Toni Morrison quotation about how she looks upon her children that particularly resonated with me, “Let your face speak what’s in your heart. When they walk in the room my face says I’m glad to see them.” So much of the time when we see our children (or our partners, coworkers, friends even!) we likely reflect what needs to be done “where are your shoes? Do you have your water bottle? Where are those TPS reports?” but children (and in fact all of us) look to see that that the person on the other side of the table is genuinely glad to see them. That is what fills up their hearts and allows them to flourish. So think about how you show up. Bring your gladness and watch how people thrive.
  3. The feeling of not belonging carries a mental load. Mrs. Obama shares her experiences being an “only”: the only black woman at the law firm, the only black woman in a class at Princeton, and how exhausting it is to not be in the “club”. I definitely cannot claim those same “only” distinctions, but I have definitely been the only woman in my class or in many decision forums and discussions and it is a different load to bear (particularly earlier in my career before I became more comfortable with myself.) I took the stories she shared as a reminder and a call to action to continue to remember the value of inclusiveness, welcoming all, seeking input regardless of communication style, and to be the change we want to see.
  4. Competency is the other side of fear. The book contains a fabulous discussion around decoding one’s fear and advice for how to manage one’s need to step up in difficult situations. Don’t be afraid; find ways to be prepared. This advice was also given with the recommendation “don’t make decisions in a period of fear; make decisions from strength.” I have always tried to ensure I’m running toward an outcome I seek when making my decisions versus running away from situations I don’t like. When you choose a path (even if it is difficult along the way) you are far more engaged and likely to succeed.

For me, the lessons the book reinforced were fundamentally about self-awareness and self-care. Your emotional state and your bias effect your decision making, and you cannot show up for others when your cup is empty. This book club session was particularly rewarding because 90% of the participants said they likely wouldn’t have picked up the book if it weren’t for the book club. I like when we have an excuse to get out of our comfort zones, so that made me particularly happy. We ended making a commitment to each other about what we would like to focus upon to fill our cups: friendship, parenting with joy, self-care in times of difficulty, making decisions from a good place, having empathy for one’s self, speaking up when you are not the majority, and being the change we want to see while holding ourselves with grace. If you haven’t read it, I highly recommend it. The book is like a big warm hug, and a kick in the pants to be your best self all at once.


A new era of creativity

I am a runner, and an early riser by nature, so getting out to run a trail at sunrise is just about the happiest I can be. The sunrise was gorgeous this morning, and I found myself meditating on the creativity of science, engineering, mathematics…we think of poetry and art as the creative arts in our society these days, but when one thinks of the history of these “hard” disciplines they are no less creative. Cuvier had to invent natural history from studying the earth and her bleached-out fossils, Mendeleev invented a methodology to organize the elements of our planet helping us predict the existence of things which had not yet been found, Euclid formulated universal truths about the relationships of physical bodies and then we as humans expanded that through observation to an entirely different field of orbital dynamics and ultimately sent people into space…these are creative arts as much as mathematical truths. We teach these disciplines as truths and forget their inventive history, which I believe discourages the very mindset that breeds engineers to be their most effective selves. For me, I only really began to learn and love my profession in the practice of it. Through working I have seen and been a part of solving some of the most exciting problems in the world of hardware, and I am confident science and technology will continue to solve some of the most difficult problems of our world (global warming, cancer, etc.) through creativity and ingenuity if we can continue to engage and unlock our future technologists.

I guess I started thinking about this because I have heard so many dystopic conversations of late where folks are worried that ChatGPT and similar AI models will displace humans doing complicated knowledge work. Why learn to code? Why figure out circuit diagrams? All of this will be automated! Me being me, it is likely no surprise that I believe these models will unlock even greater creativity and autonomy for humans through automation of the increasingly complex tasks our mounting levels of abstraction are requiring. Code still does and will always matter, being able to understand the physics of how a pump works will always matter (to a doctor trying to understand the human heart and a systems engineer trying to understand the optimal approach for liquid cooling, whether or not surgery or servicing of those servers are conducted by human hands or robots). I fundamentally believe this is the most interesting time to be alive for a technologist (and arguably a musician, artist, teacher, student, etc.) The technology that is being built today has the capacity to unleash human potential in a way formerly unimaginable, and we are the lucky ones who will see what universal truths and opportunities it will create.

I’ve also been thinking about the power of human potential and how work is or is not serving it because I recently finished Drive: The Surprising Truth of What Motivates Us. While I’ve read a lot of pieces in this vein before (Designing Your Life incorporates a lot of the aspects of flow and how you can optimize for that in your career, and several parenting books I’ve read have hammered on the disincentive of “pay to play” scenarios since they decrease the natural curiosity and propensity to learn that children are born with in favor of “unlocking the reward” behavior), I found revisiting these concepts given the current economic conditions, readjustment happening for workers in technology, and rise of more powerful AI models in our world particularly interesting. Some of the key takeaways from the book is that an incentive leads to short term results: e.g. “eat your vegetables and you will get dessert”, but less positive outcomes in the future: e.g. they don’t learn that vegetables help them feel healthier and grow stronger vs. eating dessert, which tastes good, but may ultimately make them feel sick if they eat too much, or even feel out of control negatively impacting relationships. One of the analogies used in the book is if you pay your kid to take out the trash, they will never do it again unless they are paid. It is good to have an allowance to help kids learn how to manage budgets and save. It is good to have kids do chores to help their family out of a sense of belonging and contribution to the communal needs of the family. If you correlate their chores to their allowance, then they learn to only do chores for money in perpetuity, which likely won’t help them much when they have to build their own homes and families.

This thought then make me think about my journey into engineering (the natural wonder of building my first server and playing with robots and gadgets as a kid, and how poorly my formal education fostered that same wonder), and then how my experience in work actually made me fall in love with technology and learning all over again through mentors and advocates who enjoyed what we were building. Honestly the worst parts of work have been the rote activities (writing verification tests with walking 1s to make sure there were no sticky bits–THESE TESTS SHOULD BE AUTOMATED! And then I moved to a company who built tooling to do exactly that!), and it is the complex problems, and puzzle pieces “clicking into place” where you know you are on to something and you make it WORK that make work fun. If we believe that AI will help automate the rote tasks in knowledge working (not the creativity of thinking through the problem and formulating the algorithm, but maybe the scripting to ensure that you are accessing all the right databases for the right information) and continue to expand the capacity of humans to find those big picture “aha” moments, quite possibly the future will be brighter for the individuals working to solve these problems than ever before.

Book Club

January Book Club

For our January Book club session we read Crucial Conversations. I was first exposed to this book over a decade ago as part of a training session on communication skills at Synopsys, but it was good to revisit it since I find what you take away from this book often depends on what you are dealing with when you are reading it. Since I already had the book from my previous session, I re-read the first edition, which has mostly US-based examples, and a lot of recommendations about in-person communication for the most important discussions. My understanding is that they have added content around different cultural backgrounds and remote work conditions in subsequent editions, so I’d highly recommend editions 3+. That being said, fundamentally the skills are the same: recognize when you are being triggered, ask yourself questions to help transition into an intellectual place vs. a reactionary one, and don’t get caught in a silence or violence cycle. To stay in dialogue you need to maintain psychological safety: you have to focus on what you want to get out of the discussion and use techniques to get back on track (if you see negative reactions, try apologizing for making the other person feel unsafe, use contrasting techniques to help them understand what you were trying to say vs. what they seem to have reacted to, establish mutual purpose for the discussion, etc.) I think the value of the newer editions is really around sensitizing yourself to reading queues of others who are not the same as you. Obviously techniques differ based on the medium of interactions as well as cultural backgrounds, so keeping that in mind is incredibly helpful.

On the Crucial Conversations website there are seminars as well, and of course experiencing this as a class with a cohort is probably the most ideal methodology to practice the teachings. Reading is great, but we learn through application, so especially if this is one’s first exposure to these approaches, a class would be great.

My biggest personal takeaways from the book were:

  1. Beware the stories you tell yourself. Start with “why would a reasonable, rational, and decent person act in this way?” This question helps me understand what stories I am telling myself vs. focusing on the facts. We often observe an action and assume a meaning, but in reality there are other interpretations of a person’s actions. Rather than reacting to my story, start by questioning why a person has entered an agitated state, or why you yourself are getting agitated. If you figure out how things went off track, then you can try to reset. If you get caught up in your own reaction, you cannot reestablish trust and safety.
  2. Engage in active listening–make sure I am hearing their words vs. thinking of my response. If you were clear on the goal of the meeting before the session, then the most important part of the discussion is listening to the questions, concerns, data and insights from the other parties. If you are thinking through your response or rebuttal, you are not actually listening. Never forget the goal is connection and dialogue, not a pithy response.
  3. Seek input. Especially in positions of authority, don’t speak in absolutes, which will hinder those who are more hierarchical in their disposition/culture to engage and lead to withdrawal from dialogue. As a leader, offering your opinion first almost inevitably sends the message that the decision is made, and really why are you coming to the team if that is the case? Sometimes tops-down decisions have to be made (and if that is the case, then document why the decision was made that way and don’t waste folks’ time), but if the goal is to get alignment with other stakeholders or solve a tricky problem where you do not know all of the challenges or risks, then seek input first and foremost.

I highly recommend this book, as much for discussions at home with your family as with your coworkers. When I first read this book, I was not yet a parent, and this time around I found many of the examples I thought through in my head were with respect to negotiations with my children than with my coworkers. Ultimately, when we assume the good intentions of the people around us, and show up with an open mind, we all can accomplish so much more. Happy Reading!

Our next book for book club is Michelle Obama’s The Light We Carry, which really is an acknowledgement of the times we are in, and how we manage ourselves, inspire others, and keep moving forward in times of uncertainty. I hope you will join us.


Strength Through Hope

About this time last year, I remember a conversation with my sister. She was worried about my nephew–he was feeling nauseated regularly, vomiting daily, and it was not improving. He had been through a battery of tests (he started complaining and having symptoms in November of 2020), but by January of 2022 there still was no clear diagnosis. This wasn’t our first conversation: we had been on the journey with them for nearly a year and a half as he lost weight despite concerted efforts to healthfully “bulk up”. Then he was starting to have regular headaches, issues with double vision, and all of that reinforced her conviction that something was very wrong with her son. She had been sent home from the doctor’s office regularly, but by the end of February he had trouble seeing out of his right eye, which lead my sister to the Nurseline of her insurance, and finally an Ophthalmologist telling them to go to the ER as soon as possible. There they finally diagnosed him as having a medulloblastoma, stage 4, and they had to do immediate surgery to relieve the pressure in his brain.

The next 10 months are a blur…brain surgery, PICU where he stayed for nearly two weeks, and we didn’t know if he would make it, then a medivac to Rady Children’s Hospital in San Diego where a second surgery would more completely remove the tumor, which had spread from the cerebellum to the brain stem. This led to my sister and her family applying to enter St. Jude’s study for medulloblastomas, into which Jamie was accepted, and temporarily relocating to San Diego for his treatment and recovery. He went through two rounds of proton radiation, rehabilitation, and then four rounds of intense chemotherapy.

All of that culminated in a blessed visit with all of us together in Sacramento for the holidays, from which I got home about a week ago. We did our best to make our family traditions happen: big meals around my parents’ table, cuddling on the couch to watch holiday movies, going on some outings together, and just generally trying to find a bit of normalcy in a world that makes so very little sense.

Then a few days ago we got word that the treatment was not successful. We don’t know what comes next yet. The pain of knowing everything my poor nephew has suffered was not successful in getting him to remission is devastating, but I find my most deep and profound feeling is anger at how unfair this situation is. As with all things in my life, I find myself running and reading to process…running out my anger, reading about options for my nephew, ways to provide support and comfort to my family, healthy techniques to address the feelings I have. I, like most engineers I know, intellectualize my life: ‘x’ happened because of ‘y’; therefore the key learning is when ‘y’ occurs again, step back, question assumptions, think through the outcome desired, and attempt to do ‘z’. Sadly, I don’t have a great ‘z’ for this situation. I can be angry at my sister’s healthcare provider for not taking my nephew’s symptoms seriously, not diagnosing him sooner, etc. but none of that makes this situation more bearable. It is an outlet maybe, but not a solution.

There is in fact no solution or reason for why and where this terrible disease strikes. Despite all my general positivity there is no silver lining that I can see, which is not to say that I lack hope, love, and gratitude for all the blessings in my life including this precious time with my nephew, but there is no fairness or reason that I can find, and I am angry. To cope, I find myself meditating upon resilience and acceptance. Acceptance for my anger and my inability to fix things, and resilience to be there for Jamie and my family. To build that resilience, I read, I write, and when I cannot do either any more, I run. The books I have found most comforting of late are:

If you have other recommendations, I’m all ears. I recently read this article on the psychological basis of hope and found this quotation particularly resonant, “According to these theories, hope is related to goals, anticipating obstacles, acceptance, self-worth, social support, and finding meaning in your situation. However, it’s important to note that hope must be rooted in reality, for fear of being false hope. False hope is denial. Hope itself is simply determination.”

I don’t want to have false hope, and walking that line between reality and denial is hard. I find myself determined to believe that there is meaning in this even if I don’t yet understand it. I’m just running, reading, reflecting, and seeking to understand what the greater meaning is.

Book Club

December Book Club

For our inaugural book club we read Will Larson’s An Elegant Puzzle: Systems of Engineering Management. For those familiar, Will has an excellent blog focused on the challenges and opportunities of being an engineering manager at high growth companies, and I’ve found his insights particularly useful at this phase of my career.

Since I’ve been reading his blog for a while, I was ok with the fact that this book had a lot of lists/reference tools rather than a lot of narrative, but I think it is a fair critique that from about chapter 4 onward the book feels heavy on references and lighter on the analysis of why a particular approach works, or stories of where it worked specifically in his career. I also think rules of thumb that work for him are generally applicable to the high growth market he has been occupying and are not as generalizable to larger/slower growth companies, or during periods of extreme turbulence. Still, the book provides real tools/methods for approaching problems with a concrete thesis, and to me this is a strength and not a weakness. He isn’t an academic or a researcher, so he isn’t giving the background and context for why these approaches work (I think it would be even more powerful if he joined forces with someone from that domain to make his analysis more generalizable in context), but he is clear that these are his lessons–in some sense his reflection and offer of mentorship to the world. He also makes references to others who have written academic texts in key domains worthy of further study that helped him formulate his thesis and approach (systems thinking, developer velocity, strategy, etc.) and I for one will absolutely be reading his references.

If you are an engineering manager or a manager of managers in this domain looking to level up your skill set in key domains from executive presentation, leadership, hiring, managing metrics and goals, organizational debt, and planning a reorganizaton to product management basics, I’d highly recommend the book.

For our January book club we are reading Crucial Conversations: Tools for Talking When the Stakes are High based on our poll. I have read an earlier edition of this book, but like many others in the group, I felt the learnings were critical, so it will be great to read again with a cohort. I honestly love revisiting good books: great opportunity to see how I have changed and where the narrative resonates now.

Book Club

Book Club

I’ve decided to start hosting a book club since I love reading, and would like to be able to connect with folks who have different perspectives about their key insights. Officially we will start in 2023 (part of my New Year goals to not just read for me, but to build and strengthen community). Please sign up here if you are interested.


I will publish the book as a post here on my blog that I intend to read toward the beginning of the month, and then I will schedule a virtual meet up on the last Friday of the month at 3:30PM PST to ensure that we have an hour to talk about key takeaways/findings. I will start this month just as an experimental effort to work out the kinks, so if you would like to join early, please feel free to respond here or on LinkedIn.


For December’s experiment, I’m reading An Elegant Puzzle: Systems of Engineering Management by Will Larson. Will Larson is a prolific writer, with an amazing blog that likely has advice, or at least insights, about almost any situation you might encounter in engineering management. Having managed different kinds of teams, I will say definitively that there are unique challenges to every one, and being able to step back and think about the specific group you are managing, their challenges and strengths, is a requirement. Even if you are in a position of influence with that type of team versus directly overseeing their work, it is useful to dig into the kinds of processes and flows they use, so that you can be effective in partnering.

Thanks all, and Happy New Year!


OCP Global Summit 2022: My Key Takeaways

Two weeks ago marked the end of the OCP 2022 Global Summit at San Jose Convention Center. I personally had an incredible time connecting with the community, but OCP isn’t just about connecting, it is really about having the opportunity to see technology trends across leading consumers and producers of Silicon and Systems: what is important today, and what they are building for tomorrow. Given that, I wanted to write up the key takeaways I saw (with the caveat I could only attend so many sessions). There are many great sessions which will be released after the fact by the OCP Foundation, and when I find awesome nuggets, I promise to write those up, too.

So starting with key announcements and data from the Hyperscalers:

  • Meta’s Alexis Bjorlin spoke about contributions in the domain of a new rack specification (ORV3), needed due to the thermal design requirements of AI (GPUs and ASIC-based solutions) and higher power CPU solutions forthcoming. They also contributed their next AI system specification, Grand Teton, which is 4x the performance and 2x the network bandwidth of their previous Zion EX solution, and their storage platform for AI, Grand Canyon, which offers security, power and performance improvements, and uses HDDs for AI! I followed up with some folks, and the HDD usage is for main AI servers–they are not using HDD for AI training or inference work (which was how I originally interpreted it–thanks to Vineet for helping clarify!) Grand Teton it is primarily SSD/Flash. For those listening in the room, the key nugget was definitely that AI performance, at least for DLRMs at Meta is gated SIGNIFICANTLY by network I/O. If we want to “feed the beast” we need faster network solutions (which isn’t necessarily a fancy fabric–could be standard Ethernet with better congestion management and control) and likely (eventually–no timeline specified) optics to the node to manage the power footprint of such high bandwidth SerDes. Alexis used to run the Silicon Photonics division for BRCM and Intel before that, so no surprise that optics is on her mind, but this was an impressive case study for where and why we will see the future of AI requires better network design and management.
  • Dr. Partha Ranganathan of Google spoke passionately about the inflection point in our industry (at one point saying “a career making time”) where the rate of cheaper/faster systems are slowing just as computational demand is increasing due to cloud computing, machine learning, data analytics, video, and a more massively intelligent IoT edge. What we have done historically cannot achieve the scale we need, and it is an exciting time that will require leaders to come together. He spoke about 4 major contributions swimlanes to OCP: 1. Server: DC-MHS done in conjunction with Dell, HP, Intel, and Microsoft, this contribution helps ensure we build modular building blocks for every server to maximize reuse, design isolation, supply optionality, and reliability standardized management with OpenBMC and RedFish contributions; 2. Security: a standardized RoT implementation (Caliptra, donated in collaboration with AMD, Microsoft, and NVIDIA) this reusable IP block for RoT measurement is being hardened in the ecosystem actively to ensure chip-level attestation at the package or SoC is done effectively; 3. Reliability: Along with ARM, ARM, Intel, Meta, Microsoft, and NVIDIA they are leading the effort to create metrics about silent data errors and corruption for the broader industry to track. Google is contributing execution frameworks and suites to test environments with faulty devices. I have spoken about the DCDIAG effort before when I was at Intel–this is an important approach for the industry take as complexity rises–better design and management of aging systems requires automation and testing the same way tune-ups and proactive maintenance occur on cars; and 4. Sustainability: specifically how Google is sharing best practices with OCP and the broader community to standardize sustainability measurement and optimization techniques.
  • Microsoft’s Zaid Kahn spoke on similar topics to Google (given that Caliptra and DC-MHS are contributions they coauthored), but they went even further and more specifically focused upon the future of Security for Open Compute projects. They announced Hydra, a secure BMC SoC developed with Nuvoton, which enables fine-grained control of the BMC interface authorization, so only trustworthy devices can be granted a debug interface and the access is temporary. They also announced Project Kirkland, which demonstrates how using firmware only, one can update the TPM and CPU RoT in a way that prevents substitution attacks, interposing, and eavesdropping. On the topic of modularity, the Mt. Shasta design was contributed, which is an ORV3 compliant design that supports high power devices with a 48V power feed, and supports hot-swappable modules.

In terms of manufacturers, whether Silicon or Systems, the theme was Sustainability. Samsung spoke about their renewal energy and water reduction goals. Intel and the major OxMs (HP, Dell, Mitac, etc.) showed up in modularity (leadership roles with DC-MHS and DC-SCM) and open system firmware to ensure circular economy/second life of servers can exist, and reduced embodied carbon/amortization of it over a longer period of time (given that fewer server/system components require upgrade when new CPUs, memory technologies, etc. come to market).

There was a lot of great discussion and debate on the future of AI fabrics (started by Alexis), and Ram Velaga was quite eloquent in his advocacy for Ethernet as the fabric for HPC and AI, bringing in the brilliant Dr. Mark Handley to speak about innovation in congestion management on Ethernet to unlock best-in-class performance. There was a fair amount of push on the interrelationship between compute, network, and storage for different workload scaling, and some poking at proprietary solutions addressing this (Inifiband and NVLink specifically, which seems fitting at a conference which cohosted the CXL Consortium and clearly advocates for coherent memory/accelerator pooling to move to open standard interfaces). Finally there were several sessions on optics and innovation in packaging (from Broadcom, Marvell, and Intel) demonstrated in person at the Celestica and Ragile booths, which again reinforces this attempt to use open standards to drive innovation so vendors don’t make big investments on bets that won’t have market alignment.

Conversations were also vibrant about the expansion of alternate architectures to x86 in the server ecosystem (teams from ARM, ARM-based server-class CPUs, and RISC-V server-class CPUs all were there), and Open Networking solutions (specifically a large-scale SONiC Workshop at the event). The feeling I got from collective sessions was that SONiC’s time has come, and while it still has a long way to go for feature parity, optimization, and usability enhancements compared to proprietary NOS options, the partnership with Linux Foundation for a more open and agile contribution model puts SONiC on the right track to real adoption in the industry. On the alternate architecture point of view, I feel a certain amount of conflation is occurring. Out-of-order execution exists on ARM and x86 products, so do SIMD execution units and branch prediction. From a core perspective, there is a lot more that has converged between RISC and CISC architectures since the time when RISC was introduced to the world, but where I think particular implementations have shone is the focus on the uncore portion of server-class CPU to be power efficient. This innovation is where we see real divergence in certain players (e.g. Ampere). There are a lot of things I personally still want to see in the “many cores, more power-efficient architecture” processing units (personal frustration has been that security is an “upsell” on top bin parts vs. being available across the sku stack–to me this is like saying one only gets a key for a Lexus because Toyota’s are too cheap to warrant them…ummm…tell that to the person getting their car jacked.) Security has to be ubiquitous for consumers to trust their providers, and hardware companies need to view these solutions as foundational, without significant performance impact as a foregone conclusion.

Anyway, those are my quick notes from the edge…it felt great to be there with everyone, see the innovation from Chiplets to cloud service models, disaggregated memory to further hardening of server and CPU root of trust technology. In my biased opinion the future is open, not because innovation is slowing, but rather because the complexity of the problems facing our world involve collaboration–we cannot solve sustainable datacenter design, security, performance, and reliability in isolation, and the generations to come are relying on us to succeed.


OCP Global Summit: What Not to Miss!

We are a week away from the OCP Global Summit, and I cannot tell you how excited I am to see you all in San Jose! I have been asked by a lot of folks what I think are the most exciting sessions/themes/content, so I wrote up my cheat sheet on what shouldn’t be missed at the event based on the current schedule. If you’ve never come to an OCP Global Summit, please consider this an unofficial guide!

Tuesday at 9 am is the kickoff for OCP with keynotes from industry leaders. This year is chalk full of fantastic speakers and topics as usual. You’ll get the chance to “meet” OCP’s new CEO, George Tchaparian, hear from Zane Ball on Intel’s reimagining of modern data centers, learn from Meta’s Alexis Bjorlin about open, large scale AI infrastructure (which is going to be amazing!), and check out more about sustainable computing from Ampere’s Jeff Wittich. Then you can hear from Broadcom’s brilliant execution machine, Ram Velaga about the fabric of high performance computing (hint: it is Ethernet), you will hear from my wonderful board compatriots Partha Ranganathan from Google about system design in an open innovation ecosystem, and Microsoft’s Zaid Kahn about advancing security, efficiency, and innovation at cloud scale. Finally you will hear from Samsung’s Sanjeun Cho on memory in action and the sustainable investments they have made to advance the leading edge of manufacturing technology. This is a star-studded line up of some of the largest consumers and producers of Silicon and Systems in the world. You absolutely should not miss it!

In the afternoon on Tuesday the heart of Open Compute begins. For most conferences the keynotes are the main draw, but at OCP the community and tech talks are where it is at! In our technical sessions the community shares how open standards can meet collective market needs through base specifications that allow for interoperability, reliability and scale, but still enable innovation and differentiation. These sessions basically break into four main themes: system contributions, sustainability (frameworks, thermal design innovation, etc.), open networking (based on SAI-compliant ASICs running SONiC and including such topics as SDN enhancements, monitoring, test frameworks, high performance packet processing, using kubernetes as a management plane, etc.), and memory and storage innovation (CXL, tiering, NAND density enhancements). The ones I personally am most excited about are the Cloud-Optimized Silicon for NVMe and CXL from Marvel, the SONiC workshop, and the SONiC in SDN Environment session from Google (where PINS is now supported, and this talk gets into the configuration, deployment, and monitoring enhancements required to operate SONiC at scale). 

Starting at 5 pm there is an opening night reception in the Expo Hall hosted by IBM, and it is no secret that my band, Sinister Dexter is playing. Come join us for a drink, visit the booths full of partners building on top of OCP specifications, and who knows maybe come dance with us on the floor (or bring your instrument, and sit in!)

Wednesday is one of my favorite days at the summit because it starts with the Future Technologies Symposium. OCP has two primary goals: meeting our community’s current market needs, and seeding future innovation–the Future Technologies Symposium is all about that second objective. We start early on day 2, but it is worth not partying too hard at vendor events on Tuesday night so you don’t miss the incredible volunteer leaders of OCP speaking about storage, networking, server innovation forthcoming. Some of the topics that will shine here are how we as an industry are going to come together to adopt standards for Scope 3 GHG emissions reporting, the future of software defined memory, the management of fluid in data centers (because we absolutely are seeing TDPs in processor roadmaps that are forcing more efficient use of the power we have ESPECIALLY at the edge), updates on cloud serviceability models and how we need to operate across system components to ensure reliability, security, and scalability, storage disaggregation, telemetry and workload planning for dynamic control and sustainability in data center operations, chiplet proof of concept and ecosystem advancements, optics, and much much more. The Board’s very own Andy Bechtolsheim closes out the sessions, and then there will be a Block Party hosted by Meta and Google at the main entrance plaza of the convention center to cap off the day.

Thursday is the final day of technical sessions, and I may argue the best day since one of my very own team members, Sung Park, will be presenting that day. Day 4 is more focused on manageability, security, and server modularity specifications (DC-MHS, ORV V3, OAM, OCP NIC 3.0, OSF/MinPlatform, etc.) I personally will be looking forward to the DC-MHS M-DNO concept review with Dirk Blevins and the DC-SCM 2U/wide chassis presentation with my friend Siamak Tavallaei presenting. The theme of sustainability runs throughout the day from open system firmware and management, new standards so the community can see through marketese in this domain–Eric Dahlen and DJ will be co-presenting on new standards for modern data center assessments–PUE is a metric of the past friends and I’m super excited to see the biggest consumers and producers align on how we will report emissions, modularity (to maximize reuse of server components lowering embodied carbon footprints over time and reducing e-waste), and better thermal design from the silicon and systems to dynamic operation and control. There are even sessions on Time Appliances and synchronization mesh methods for clocks (turns out this REALLY matters in global fleets–glad to see standardization around it because not every server can be shipped everywhere, and there are some problems hardware is better at solving), network performance anomaly detection, and a lot of security innovation on top of DC-SCM (Caliptra, HYDRA, etc.) for disaggregated root of trust and attestation flows to improve security and isolation at scale. 

Long before I was on OCP’s Board, I viewed the OCP Global Summit as the industry event where the community speaks about their real world applications of new technologies. We are the systems builders, the ones who have to make it real. I cannot wait to see you all there to empower open!