4. INTEL DATA CENTER
GROUP
MOVE | STORE | PROCESS
DRIVE THE BEST EXPERIENCE & SLA
PROTECT DATA PRIVACY
DRIVE HIGHER REVENUE
REDUCE DOWNTIME
DRIVE OPERATIONAL EFFICIENCY
CLOUD
SERVICE
PROVIDER
S
COMMUNICATIO
N
SERVICE
PROVIDERS
ENTERPRIS
E
AND
GOVERNME
NT
SILOED APPLICATIONS & DATA PACKETS
SLOW DEPLOYMENTS OF NEW SERVICES
SECURITY EXPLOITS GROWING
DATA MOVEMENT & NETWORK BOTTLENECKS
CONVERGED WORKFLOWS & INFRASTRUCTURE
PREPARING FOR 5G
NETWORK TRANSFORMATION
DEMANDS AT THE NETWORK EDGE
NEW SERVICE OPPORTUNITIES / RPU
5. Build the Future with Data-Centric
Infrastructure from Intel
StoreMove
SILICON PHOTONICS
OMNI-PATH
FABRIC
ETHERNET
Process
16. Customer
● Development of its application
● Application architecture and choice of
middleware
● Design of unit tests
Clever Cloud
● Maintenance in operational conditions
● Build Management
● Automated production runs
● Data backups
● Supervision and scalability
● Security audit and updates of
continuous systems
● Networking
26. • From KVM to Rust-VMM
• VMx is a straight access to
CPU full capacity using
privileged mode on a
sandbox
27. • Use it as a container for software part and containing
• New FaaS (Function as a service) using unikernel fast boot
instance
• Boot VM in less than 50ms
• Use WebAssembly (WASM): C, C++, rust, Go, Kotlin,
Assemblyscript…
29. For its Public Cloud, Clever Cloud has developed a long-term partnership with
HPE.
○ Clever Cloud Public comes standard with
latest generation HPE Synergy hardware
○ All hard drives are SSDs
● In case you want additional material resources, a
connection for preferential rates is possible
Editor's Notes
We work together on Rust-VMM Cloud hypervisor as part of our enabling ecosystem efforts
In order to fully access the value of data, we need to embrace the confluence of 3 mega-trends. The first is the democratization of AI.
The industry has shared its innovations in AI for a number of years, but examples of innovation have been concentrated to the largest data companies on the planet.
Today, we are now seeing the signs that AI is shaping the future of entire industries
The second mega-trend is 5G.
5G is revolutionary and accelerating the need for a flexible, automated network to fuel the growth of the data centric edge to deliver compute at the location of data creation.
The third mega-trend is cloud computing.
The most mature of the mega-trends, cloud computing is built on Intel Architecture and innovation, such as Intel virtualization technology first delivered in 2005. We see cloud architectures extending from cloud service provider data centers across the enterprise and core networks and to the intelligent edge.
Key Message:
Across the major customer segments, and industries worldwide, the urgency to make better use and impact of data has never been greater.
Each of the major infrastructure providers experience various challenges as they seek to evolve themselves to service this new data centric world – from edge to cloud and back.
CloudSP:
Need to ensure the most flexibility of cloud scaling is fully realized
CSPs must constantly optimize the utilization of their infrastructure to meet customer SLAs, reduce downtime, offer new differentiated services to remain competitive,
deploy tools and services to developers to help them prepare for their own datacentric needs.
CommSP:
We just concluded a very exciting week at the 2019 Mobile World Congress.
Data is at the heart of an unprecedented 5G market opportunity estimated to reach $10 trillion by 2035.
The vast volume of data generated and acted on every day is driving transformation of communication networks, disrupting industries and inspiring a global ecosystem of technology companies to seize a seemingly limitless market opportunity.
The cloudification of the network will require flexibility, reliability and security to ensure the workloads deployed on the 5G network will be able to meet the fluctuating demands and shifting SLAs at the edge.
Enterprise & Government:
For years, enterprises have been undergoing a digital transformation to shifting their infrastructure to realize the benefits of a multi-cloud world; to deliver new services faster and save on operational efficiencies.
At Intel, we have been investing in silicon innovation for years to drive the right technology foundation to move, store and process data across the data centric compute landscape
This has included ongoing investment in our CPU portfolio, but also includes leading products for data movement such as our Ethernet product offerings, delivering storage innovation and efficiency with Intel Optane DC and QLC 3D NAND SSDs, and creating unique discrete compute acceleration with Stratix 10 FPGAs.
It also means accelerating investment in new spaces including discrete acceleration technologies for AI workloads, disruptive innovation for persistent memory, and networking with our Silicon Photonics products.
Intel’s strategy is to drive a new era of data center technology innovation that enables every IT organization to benefit from their data and accelerate their successes
Very simply Data-Centric infrastructure of the future will:
Move data faster… removing bottlenecks within the data center as traffic continues to growth
Store more data, with the ability to quickly access that data to deliver rapid, real time insights
Process everything with through a combination of general purpose and workload accelerated compute products
The software ecosystem Erwan just mentioned is fundamentally based on the biggest technology revolution of the past few years: Open Source. The cloud runs on open source, 5G runs on open source. The data moves between users and hardware through a very large set of open source layers.
At Intel, open source is the key vehicle for us to enable and deliver our silicon innovations to our customers. This is why for many years now, we’ve been leading key parts of the open source ecosystem.
Take the Linux kernel for example: This is the first software layer sitting between our hardware and the rest of the software stack. The Linux kernel exposes our hardware advantage to the rest of the ecosystem. We probably have the strongest Linux kernel software engineering team in the world, and as matter of fact we’ve been on top of the kernel contributors list for many years now. We know open source because we build it.
The open source virtualization stack is another part of the software ecosystem that we lead. Virtualization used to mean full blown virtual machines, but we’re now seeing it used to build an isolation continuum in the cloud. All the way from large scale Kubernetes infrastructures down to containers and serverless functions, all those use cases are now using virtualization as a workload isolation mechanism. So we decided to take a new approach on open source virtualization to move from the current monolithic architecture to a more modern one: Rust-vmm
Rust-vmm is an open source project that we Intel, together with AWS, Alibaba, Redhat and Google created. The idea behind rust-vmm is to build independent and open source virtualization components, and then allow users to compose hypervisors that fits perfectly their use cases and nothing else.
Based on that approach, we went one step further and used rust-vmm to build a cloud focused hypervisor, with the single purpose of exclusively running cloud workloads.
Cloud Hypervisor is an open source project and it’s already being adopted by several key cloud providers like Alibaba for their public container infrastructure. With this project we’re building a safer, lighter and performant hypervisor exclusively for virtualizing modern cloud workloads. So please join the project and help us define the next cloud virtualization generation.
Possible transition: One company that’s consistently looking at and contributing to innovative open source projects like rust-vmm and cloud hypervisor is Clever Cloud. Please give a welcome its founder, Quentin Adam, to the stage.
We work together on Rust-VMM Cloud hypervisor as part of our enabling ecosystem efforts
