Soni Jiandani Introduces Pensando Systems

thank you very much uh and it's a pleasure to be here and sharing with you the news that we announced with vmware a quick background for those of you that are not familiar with pensando as a company uh we are a a group of individuals and folks that have worked with each other over the last two and a half decades and chairman of the board is uh john chambers our ceo is prem jane randy pond is our cfo and mario is a vice chairman and luca cafero is a board member and a very active one and i am the chief business officer a core part of our winning strategy has been to be very customer and partner focused and focusing on the speed of innovation it has resulted in delivery of eight major platforms spanning a multitude of industries including switching routing voice uh platforms uh unified computing systems software defined networking storage area networks to name a few the goal clearly has been is to continue to innovate with pensando and to leapfrog prior generation leaders and enable what we call the democratization of the cloud we would like to consider ourselves to be the new definition of the 20-somethings as we start to drive another major market transition in the marketplace if you take a look at how the clouds have been scaling and particularly at the hyperscalers like microsoft azure amazon and google cloud what they have fundamentally done is they have built custom accelerators these at a system level essentially allow you the opportunity to share your resources whether they are cpus memory storage peripheral devices into pools that can be operationalized autonomously and to create a set of consumable infrastructure platforms and unified apis now clearly they were working from a point of advantage with a unique business model because they started with a clean slate design and there was no legacy to protect they also came in incredibly well funded and a longer-term roi expectation ironically while a lot of these custom accelerators have been built by these companies they are not available to the general public and they are not available to the tier 2 tier 3 cloud providers our strategy essentially is to build such an accelerator and to democratize it and make it available to the cloud customers and our enterprise and service provider customers alike if you compare and contrast our currently shipping platform which we are into production at scale with numerous customers you will find and when you draw comparisons visa the aws nitro which is their custom development offering across their cloud you will find anywhere from nine times improvement around latency and jitter which essentially allows us to support edge applications like 5g and iot very effectively along with the amount of packet per second processing where we can do up to eight times the amount of packets on our programmable processor when you contrast that to the aws nitro allowing our cloud customers to maximize their revenue per rack and the ability to establish up to one million new connections per second with power usage which is one-third that of aws in the case of aws they require five times the number of devices to accomplish the same set of services that we require a single device for ultimately the goal here is to deliver this technology to all of our customers that are desiring to get the performance of hardware with full programmability including the data plane for software allowing this technology to be widely applicable across enterprise workloads cloud workloads and service provider environments now when you take a closer look at the enterprise what are the industry folks saying about what the requirements are of the enterprise as they look for this new architectural approach they fundamentally will have the ability to acc to deploy this technology at scale and leverage the same building blocks that we used that are applicable for the cloud and bring it into their own private cloud it will allow our enterprise customers to transform beyond the traditional scale-up models as they build the next generation application infrastructure and the announcement we made with vmware is another validation of that precise point many a times you will hear different nomenclature being used whether it is gartner that calls it a fact or idc that calls it a focp or nvidia that calls it a dpu now while all accelerators are essentially network interface cards not all network interface cards and smartnics are accelerated and this is what is extremely important to look deeper around and if i were to quote a colleague from gartner the expectation in the next three years is that one in every three network interface cards will be shipped as a fact out of the factories of the server vendors now when we were building this accelerator from ground up what were some of the things that were very important to us what we heard our customers telling us is that the environments of today are extremely complex and when they start to build data centers using the legacy models one of the biggest challenges they face is the is having to trombone a lot of traffic and not have the flexibility to chain all the services the power of our technology essentially can be unleashed by enabling our customers to turn on multiple services simultaneously while maintaining a very low power footprint that enterprises need particularly as they deploy different types of servers within the enterprise environment guarantee a low latency and jitter while all the services are enabled at scale and the ability to deliver it either at 10 gig 25 gig or 100 gig all of the services that we will be talking to you about are completely software defined and this is again goes back to the power of our technology whereby everything including the data plane is fully programmable allowing the customers to turn on security services at scale while turning on networking capabilities and storage services and delivering line rate monitoring and real-time telemetry to reduce their mean time of repairs and making these services in line available all the time at wire speed and a pay as you grow model allowing them to turn on those services as and where they need them as opposed to being required to acquire these assets and being forced into turning on capabilities they want we also built a policy services manager which allows our customers to manage the life cycle of these devices across their compute infrastructure from a centralized location while providing them the ability from a single location to manage the policy and the ability to deal with troubleshooting and reducing reducing the mean time to repairs everything is driven through restful apis whether customers want to consume it through the policy services manager or whether they want to deal directly and interact directly using the same restful apis when interacting with the pensandro distributed services card through the northbound apis of the policy services manager we then integrate into a broad ecosystem which you will be hearing more details around but let's spend more time now talking about the strategic collaboration between vmware and pensando i couldn't be more excited to talk about this and to introduce our colleagues from vmware as we have worked for many many years and combining decades of innovation across compute virtualization networking storage and security and as we bring project monterey which is built on a co-development model between our two teams the goal here is to deliver our customers a future-proof platform that will deliver the same building blocks with cloud-like capabilities and the efficiencies and economics to the enterprises and cloud providers and service providers and this technology today is only applicable in the hyperscalers can now be democratized between our two companies collaboration efforts customers can start the journey today with vmware's cloud foundation platform running on the pensandro distributed services cards today

Channel: Tech Field Day

Views: 742

Rating: 5 out of 5

Keywords:

Id: AFq2FDiB63s

Length: 9min 42sec (582 seconds)

Published: Thu Oct 15 2020