2019 - The Year of Hybrid Cloud
Transcript
hey it's Keith Townsend from the CTO visor 44 zettabytes what does that have to do with your data center strategy going into 2020 well let's take a look back at twenty nineteen sponsored by Intel [Music] one of the big predictions earlier in a decade was that by 2020 a year and now there will be 44 zettabytes 44 trillion bytes of data throughout the world stored on disk somewhere it's impossible to process that much data so IT teams have already been asked to do the impossible and
2019 was a step forward and kind of solving that problem what did I see as some of the major trends the biggest make train of last year I would say is the acknowledgement that hybrid IT is the future you know here on the CTO advisor we're kind of betting our practice on the simple fact that private data center and public cloud will coexist well we're seeing were close being repatriated back to the private data center because of this cost factor which we'll get into and to
this data gravity challenge that enterprises are trying to solve you can't simply move 44 zettabytes of data around from your on-premises systems B IOT om systems data center Big Data ERP into the public cloud where data is steady growing itself you need to process that data in place Dave McCrory has said that data has gravity and I agree with that data has gravity but we've moved slowly to being able to move that compute especially these complex pieces of compute where we need to manage this data
what do I mean by the industry has accepted hybridity well take no further look than all the cloud providers they've acknowledged that there needs to be a or premises version of the cloud whether it's GCP AWS azure everyone has a on-prem version we're start you see traditional data center players like VMware have a motif cloud strategy to complement their on-premises solution but where's Intel playing in this room role I think the biggest area outside of processors obviously is the ability to process data or get data
to the processor si P I think has led the way when it comes to in memory databases and promoting the power of a memory databases and the enterprise however stuffing a modern server with 1.5 terabytes of RAM or 3 terabytes a ramp you have a dual product system is cost prohibitive this is where the persistent memory that Intel has released in 2019 which has been promised I think most of the decade comes into play this persistent memory allows kind of the best of both worlds frankly
you get the lower cost of NAND flash memory with the lower latencies of that's closer to DRAM and the higher capacities so that systems like si P and all the way down a sequel standard can take advantage of this innovation I can have a three terabyte single Hana system that scales out to multiple terabytes for analytics I think the capabilities of what we can do with a memory databases are just starting to hit kind of that 44 zettabytes allenge the other piece processors it tells us
got some competition and processors whether it's arm or the other x86 providers the data center has become a lot more competitive and what I've seen Intel do in response to that competition is lean on what they do best which is scale at an AWS summit in New York AWS and Intel Co presented on the value of this long-term relationship you can go from a m4 to m5 workload and speed up performance by 25% and multifold for some of these analytic workloads AWS always talks about how
things break at scale and how important it is to engineer for the scale of an AWS provider across all the tile providers Intel has developed the chops to scale their systems now you bring that back into the data center what does that mean for the actual data center why should I stick with Intel when competitors like arm and other x86 solutions have some pretty interesting speeds and fees this isn't just a conversation about speeds in feed this is again a conversation about scale how do you
scale your operations across private cloud and public cloud and I look forward we'll talk about telemetry data the ability to get information off of the systems and systems across multiple environments and what that means to the data center but from a simple perspective Intel has continued to forge their relationships with the major providers HP Lenovo Dale all have deepened their relationships and adopted not just the persistent layer of memory the Zenon scalable processors but also the network components that wall that work together to improve overall
system performance at scale we'll talk about the potential of these systems working together in the look forth I'd love to hear your feedback do you think Intel is on the right track when it comes to meeting the needs of your data center when that data center stretches beyond just your four walls and into the public cloud what can they do better share it with me online I can be found on the web the CTO visor comm is the website CTL visors - Twitter handle talk to
you next CTO DOS
