Everything Converged at Gartner Data Center

In addition to these great sessions there's also an amazing VUG Dinner Event going on Monday Night!  Learn more and register here: VUG Gartner Data Center Conference Dinner Tickets, Mon, Dec 5, 2016 at 7:00 PM | Eventbrite . 

Agenda includes cocktail and dinner networking with many CPSD Leaders, including Chad Sakac, Peter Cutts, Glenn Keels, Bob Wambach, Jon Siegal, Michael Ficociello; a fireside chat with Chad, Peter, Glenn, and VUG Member from CANSOFCOM, discussing their successful IT Transformation and what 5 Blocks and 7 Rails has enabled them to accomplish.

Hi Everyone,

Just got out of The 2021 Storage Scenario: Planning for the Future While Deploying Today session.  It wasn't on our original list but a huge part of it was integrated systems, specifically hyper-converged systems.The big takeaways for me were:

• all flash configurations are proven and ready for mission critical applications
• HCI has superior TCO because
  • it productizes services &
  • reduces configuration issues
• 20% of mission critical applications will be run on HCI by 2020
• Larger companies have embraced HCI (duh) and a leveraging it as a competitive weapon.  Meaning that they are not afraid to disrupt themselves by moving away from 3 tier systems

Do you agree disagree? if you were at the session what were your takeaways?

I'm keeping a list of all my take away notes using Twitter Moments this year.

#GartnerDC 2014, 2015, 2016

I haven't seen Matt Baker (@MattWBaker) present before so I pleasantly surprised by the very holistic approach of the Delivering the IT Tapestry of the Future session.  The key takeaway here is that there are many different levels between On-Premises and Off-Premises that aren't getting as much attention.  On top of that there does not seem to be a golden ratio for how much of each you should use.  It is definitely dependent on your needs.

On that note there are a number of us here that you can follow or reach out to:

Chad Sakac:@Sakacc

Jay Cuthrell: @JayCuthrell

Jon Siegal: @Jon_Siegal

David Giesecke: @DGiesecke05

Paul Young:  @YoungP2`

Definitely a fan of the Networking Matters in a Hyper-Converged World session.  The big message here was that the networking teams and hyper-converged teams need to work together which was highlighted by the fact that 90% of the folks in the room were networking by show of hands.

The most encouraging part was that companies are working with partners or are integrating across owned assets to integrate these two domains.  I was kind of surprised VxRack wasn't mentioned here since this is a perfect example of that integration.

I think the biggest talking points are covered on the slides below but the take away was as many complexities will come up there are things you can do to get ahead of them

Jay explained it well

To Buy or to Build: Understanding the Different Use Cases for Converged and Hyper-Converged was well worth the standing room only seats.  Chad Sakac covered everything from Bundles and Validated systems to Converged infrastructure and Solutions.  The message that was consistent though was that people care less and less about what the product(s) is/are and  more about the business outcomes they can deliver.

I don't think I could do it justice here but thankfully Chad covered a lot of this material in his blog or you can read the recent CRN article.

A new taxonomy: the Build to Buy continuum - Virtual Geek

5 Steps On Dell EMC Hyper-Convergence Chief Sakac's Map To The Cloud - Page: 1 | CRN

Re:The impact of composable on next-generation hyper-convergence

Yesterdays session on composable infrastructure provided really good information from one of the industries strongest minds in this space George Weiss. He spent some time addressing some vendor differences addressing both HPE's Synergywith OneView and Dell EMC's Active System Manager. 

There were several key benefits discussed including simply assembled and managed building blocks, Incremental just-enough scale and capacity, dynamically integrated versus pre-integrated, enabling SDI and  SDDC and reducing  life cycle costs. Probably the biggest benefits for customers are around reducing Opex and speeding time to value with infrastructure.

While a lot of the industry call this space composable infrastructure we look at this on a much broader spectrum and talk about service defined infrastructure. Being able to compose dis-aggregated infrastructure is only a piece of the larger puzzle.  For more information on service defined infrastructure see here.

Note: Can be have shorter session titles next year?

Dissecting the TCO of Software-Defined and Hyper-converged Systems vs. Traditional Architectures had some key insights that might seem obvious but aren't things we are always thinking about.

1. With great flexibility comes great responsibility: I hope the comic book geeks like me appreciate this comment but it really is true.  The biggest difference between software defined and HCI is that  software defined gives you almost limitless flexibility BUT it puts all the responsibility on your shoulders. Where HCI is takes away those choices of what parts & pieces are used but you have single vendor support that will troubleshoot the environment not just a tiny part of it.
2. Don't compare your old infrastructure to the most modern solution: With the sinking costs of storage and management finding something better than what you had is not a high bar to hit.  As Julia Palmer said "you don't want to compare apples to bicycles"
3. HCI and Software Defined come out on top frequently but they do not solve everything: Many of the HCI vendors don't sell anything other than HCI and software defined so there are plenty of people that will say it can handle any application and any scale and it's not true.  One clear case study showed that if you have large and exponentially growing storage needs All-Flash arrays could be the better option
   

There were also 2 great slides.  The first is what to have in mind when you are comparing technology and the second is how to stay prepared for the next wave of tech whatever it is (AKA Julia's LAME cycle).

Here are my key take-aways from the HCI tech session that Paul Delory delivered. Overall, a great session from Dell EMC perspective where VxRail was highlighted as being on Gartner's "short list" for HCIS.

Some key takeaways:

• VDI top use case for HCI. (No surprise there).
• Paul then “conditionally” recommended HCI for relational databases. Best used for proliferation of smaller databases.  Cautioned that HCI data services don’t always work well with relational db’s and that relational db’s scale up and therefore often require just more storage (not CPU/RAM too) each time adding capacity. 
• HCIS Appliance is “trojan horse” for deploying SDDC management technology
• Networking a “pain point” for HCI due to huge amount of east-west traffic (opportunity: VxRack!)
• Evaluated 3 major players in HCI – Nutanix, VxRail, SimpliVity
  • VxRail
    • Strengths: pre-built, turnkey; tight integration with VMware; NW integration as ref arch; support/service
    • Weaknesses: Data services lag, bolt-on data protection
  • Nutanix –
    • Strengths: enterprise grade platform, Acropolis hypervisor is free alternative to VMware/MSFT; can do inline dedup/compression, but can turn off; Prism
    • Weaknesses: Storage performance and lacks network integration