overview 10+ locations silicon valley ca | bellevue wa | london uk | belfast n ir nw development center seattle, wa worldwide hq silicon valley, ca emea development center belfast nir emea hq london uk founded in nov 2018 by ex-3par executives 3 siamak nazari, chief executive officer 3par chief software architect, hpe fellow craig nunes, chief operating officer 3par cmo, hpe storage global marketing vp sean etaati, chief technology officer 3par platform director, hpe/3par chief hardware architect david scott, executive chairman 3par ceo, hpe storage svp/gm board: appdynamics, overops, veritas, token, couchbase engineering talent cloud, security & storage patents filed
increase asset utilization reduce costs and infrastructure footprint increase operational simplicity and agility reduce operational overhead and accelerate deployment of new services take advantage of cloud innovation across my enterprise give application owners a self-service on-prem experience
enterprise-class storage that consumes no server cpu/memory resources and is defined and managed through the cloud secure, ai-driven, cloud-based control plane called nebulon ON application server clusters with PCIe-based services processing units (spu’s) replacing the RAID or FC adapters + 5 up to half the cost of expensive enterprise arrays leveraging less costly server SSDs
application owners • have low latency, mission critical workloads on-prem • tried sds or hci, but sla or other workload restrictions forced continued use of external storage arrays • manage 10’s, 100’s or 1000’s of servers • may have multiple sites or edge locations example customers • cloud service providers: looking for new cost-effective ways to deliver value- add services to clients • mid-to-large enterprises: under pressure to reduce costs and offload the overhead on operations staff
vendor 7 major server vendors will offer cloud-defined storage • founding partners: hpe, supermicro and one other major server vendor • spu, subscription available as a server option sku from each partner • all quotations and order fullfillment provided by the server vendor • all support agreements provided by the server vendor
the application server 9 • runs in any 2u/24 drive server • connects to server ssds like any raid card • presents local or shared volumes to app • group of spu-equipped servers: npod npod servers spu storage media o/s or hypervisor boot vol data vols
all flash array within a card… 11 hypervisor/os agnostic no software required on the server all-flash array latencies powerful cpu’s/processing 1.33x app/vm density advantage vs ‘restrictive’ sds separate fault domain data/drive access even if server fails petabyte-scale npods up to 32 servers/pod medusa spu enterprise data services dedupe, compression, erasure coding, encryption, snapshots, data mirroring workload aware using tailored application templates …plus half the cost of ‘expensive’ external arrays
kubernetes, openshift cloud native bare metal nosql, message brokers containers virtualization server virtualization with vmware, kvm, hyper-v bare metal single instance or clustered sql databases
redundancy techniques shared, non-mirrored for ephemeral apps requiring shared storage shared, mirrored for mission critical apps requiring shared storage local, mirrored for mission critical apps requiring single writer access to storage automated storage configuration options for nPods kubernetes a k8s node accesses highly available persistent volumes on a single host mongodb multiple hosts access local storage with data redundancy at the app layer vmware development a cluster accessing shared storage, don’t require storage for redundancy vmware production a cluster accessing shared storage, and rely on storage for redundancy
• always up-to-date; as-a-service • make use of new features instantly • enriched with aiops capabilities services processing unit • lightweight storage OS: nebOS • fast, infrequent updates with no disruption • scheduling updates is easy management from the cloud control plane 16 management stack I/O stack I/O stack I/O stack traditional storage operating system ~25% I/O stack ~75% management stack
infrastructure disaster tolerant • multiple availability zones so management is always available for any infrastructure scale • zero ops impact if cloud connectivity lost role-based access control • all actions are secured and audited to allow and monitor delegated management data is encrypted • in-flight and at-rest for user interactions, spu’s, and cloud services to keep all data private security triangle • secure tokens so only authorized users within the firewall can make changes spu user browser spu cloud api user cloud api token delivery authorizes the execution of changes private key firewall sends config/status & receives instructions queries the cloud & issues changes to spus/npods patented ‘security triangle’
storage, server & application telemetry ai-assisted administration • AIOps for predictive maintenance & faster troubleshooting • zero touch provisioning api-driven automation • control all devices from a single url in the cloud • automated software updates & fleet management chicago atlanta phoenix london berlin global infrastructure insights • 10’s of 1000’s of server, storage, & app metrics / hour • versatile reporting drives better decision-making
19 • curated, reusable definitions of server-storage configurations for specific applications o vmware, kubernetes, mongoDB, more to come • repeatable, consistent infrastructure deployed at scale • template automation includes: o volume number & size o data redundancy & protection o snapshot schedule o server accessibility to volumes o os image for boot volumes • constantly optimized using system telemetry
20 volume management becomes app provisioning • app templates provision what the app needs • app owners get actionable recommendations from ai-ops • scaling is as easy as adding (heterogenous) servers shared workloads becomes workload awareness • app owner owns the full infrastructure stack • per-app npods allow for workload aware data services querying multiple systems becomes one api • full visibility across the infrastructure • better control of app policy compliance
expense of array-based storage without suffering the restrictions of SDS/HCI • enable self-service for application owners and reduce the burden on infrastructure managers • available from your existing server vendor
itpresstour
spatial_ai_network
fossamagna
mhrtech
ysknsid25
masuda220
news_it_enj
kaminashi
nari_ex
quiver
genda
glaforge
smt7174
twada
trishagee
3n
sstephenson
deanohume
apolaine
jacobtomlinson
livdayseo
sugarenia
bluesmoon
jonyablonski
rabernat
