Dave Hitz, NetApp Inc.’s executive clamp president, was one of a vendor’s founders in 1992 and
has had a front quarrel chair to a raft of changes in a information storage attention for 20 years. We
recently held adult with him to speak about NetApp
storage strategy and attention trends, including clustering, converged
for virtualization, primary
deduplication and cloud
SearchStorage.com: What is your stream purpose during NetApp?
Hitz: My purpose is arch gadfly, presumably visionary. we started as a programmer. I’ve worked
as outbound product preacher and ran engineering for a while. we took a year off and wrote a book, got concerned with
transition formulation from [CEOs] Dan
[Warmenhoven] to Tom [Georgens], and started essay papers called destiny histories.
We’ve finished a series of vital transitions of a company. In a mid- to late-’90s we took a big
bet on a Internet and grew adult as a company. After a pile-up hit, we re-oriented again and based
ourselves on virtualized applications. Then we placed a genuine large gamble on server virtualization.
Every 5 years we’ve re-invented ourselves as a company.
The gamble we’re holding this time is on large-scale infrastructure for data, in particular.
SearchStorage.com: What purpose does storage play in this large-scale infrastructure?
Hitz: The universe of information has altered a lot over a past few years. Ten years ago, it was
mostly tedious corporate things … schematics, payroll, budget. Now we’re articulate about storing baby’s
first steps, letters to a spouse, medical annals — deeply personal stuff. So a space of data
is unequivocally changing.
The other thing that’s function with information is a place where many information government used to
happen has disappeared. Data government used to mostly be on a earthy server. These days with
virtualization, a earthy server is not unequivocally many of a place. The practical machines are
flitting from one earthy server to another. And a practical servers aren’t a good place for data
management given typically we have mixed practical servers pity information joining into the
same database. So a doubt is ‘Where does a information government go?’ And increasingly, it’s been
moving into a storage layer. The outcome of that is storage is starting to spin an
infrastructure for data. What we meant by infrastructure is that it’s a apparatus that’s shared,
used by lot of opposite people nonetheless managed centrally.
SearchStorage.com: How is this reflected in NetApp’s
strategy for a storage?
Hitz: A large partial of that is a clustering we’ve expelled in Ontap 8. We consider clustering
is unequivocally foundational to be means to have these characteristics where we can take out a server and
install a opposite server. How do we make that occur when there’s information on that storage server
that needs to be there? The answer is we need to have these things clustered together with the
ability to only transparently quit information from one place to another.
SearchStorage.com: NetApp has worked on clustering for utterly a while, unequivocally given acquiring
Spinnaker Networks in 2004. Data Ontap 8 was ostensible to be a chronicle with full clustering, but
you didn’t speak about that many when 8.0 came out. And we still charity Ontap in Version 7 mode,
without clustering. Why have we waited until Ontap
8.1.1 to unequivocally stress clustering?
Hitz: When we shipped 8.0, it was a flattering radical ascent of Ontap, kind of
corresponding to when Microsoft came out with Windows NT. we mean, large architectural changes. We
took a opposite proceed — when Windows NT came out there was a classical mode, nonetheless that wasn’t the
default. When we came out in 8.0, it ran by default in 7 mode. That means we did this radical
architectural ascent with all a foundations for clustering in place nonetheless didn’t boat it with all
that things incited on. And we didn’t make all that large of a understanding of it during a time given we wanted
to get all a pieces prepared so that people would be means to flip it on.
So 8.1 is indeed a place where we’re revelation a story of clustering, when we’ve got proof
cases commissioned now and business regulating it. So we took some-more of a ‘Let’s get a infrastructure in
place, get people regulating on it and afterwards make a lot of sound [approach].’ So here we are in 8.1 and
we’re revelation a story of 8.0. Maybe we’re too regressive [in our] marketing, nonetheless we like to
know what we have before we blow a bugles.
SearchStorage.com: How many NetApp
storage business are regulating clustering?
Hitz: It’s a comparatively tiny percentage, singular digits. Our idea was initial to get 8.0
deployed with a whole new architecture, and make certain that whole new architectural horizon was
SearchStorage.com: While we were operative on building clustering into Ontap, EMC
bought Isilon and had a mature clustering shipping product. How does NetApp’s clustering
stack adult vs. Isilon?
Hitz: You have a unequivocally tough choice if we go with Isilon. Because on a one hand, EMC
has a lot of engaging capabilities: snapshots, cloning, dedupe, compression, a abounding set of data
management features. The problem with EMC is that for any of those facilities we have to contend ‘Wait
a minute, that product has it?’ And it’s not uniform opposite them. You contend ‘I wish to use SATA
drives.’ Well, we can do that for Isilon; if we wish to do this other thing, it’s VMAX
and [for] this other thing it’s VNX.
If we demeanour during facilities EMC can support, we finish adult with a finish list. If we mangle apart
their architectures and demeanour during a same underline list by architecture, we finish adult anticipating a main
feature Isilon has is clustering, that is great. Unfortunately, it’s not in multiple with the
full apartment of abounding information government capabilities. That’s a No. 1 disproportion Ontap has — it’s the
same Ontap that has all this cold things in it.
Second, what people find with Isilon is that we can scale it, nonetheless when it comes time for a
hardware refresh, you’re articulate downtime. You don’t take down your network when we put a new
router in place. You shouldn’t have to take down your new information infrastructure when we put new
storage systems in place.
SearchStorage.com: A lot of IT vendors are relocating toward a converged infrastructure combining
storage, networking and discriminate into one offering. NetApp does this by a FlexPodreference
architecture. Will joining be a widespread trend in storage implementation?
Hitz: There are dual competing trends. we have this design in my conduct of a smoke-stack of
stuff we need to put together a information center. There’s a CPU chip itself, a handling system,
applications like databases, capabilities like networking and storage. You can consider of that as a
vertical stack. On one palm there’s vigour to get that whole straight smoke-stack integrated to make it
easier for customers.
There’s another trend, horizontalization, with companies meditative of their possess sold layer.
When we consider of that trend, we consider of companies like Intel with a chip or Microsoft with the
operating complement or Cisco with networking. The doubt is that of these trends is winning? It
looks to me, in terms of IP tenure and a tangible record development, a plane model
is winning. Oracle’s database beats any of a aged server guys’ database; Cisco’s networking beats
any of a aged server guys’ networking; and EMC, Hitachi Data Systems [HDS] and NetApp are
two-thirds of a storage market.
SearchStorage.com: You discuss EMC, NetApp and HDS winning a storage market, nonetheless there’s
still a far-reaching opening between we and marketplace personality EMC. How can we tighten that gap?
Hitz: Ontap is a No.1 storage design in a market. EMC is still a good bit
larger than us in storage marketplace share, nonetheless EMC is deploying lots of opposite architectures.
They’ve got Enginuity for Symmetrix and apart architectures for Clariion/Celerra, Data Domain
and Isilon. If we demeanour during marketplace share by architecture, NetApp is No. 1, EMC is No. 2, 3 and 4,
and afterwards NetApp is No. 5 with a E-Series.
Even nonetheless EMC is offered some-more storage, it’s not a best building retard for infrastructure
because we always have to stop and say, ‘Wait a minute, did we wish it to be Celerra style, VNX
style, VNXe style, VMAX character or Isilon style?’ All of those are opposite things.
We have a E-Series for impassioned cases. We contend a E is for Extreme — super-high bandwidth,
Teradata-style storage. But in general, we tell them, ‘Ontap’s a answer.’ It’s a simple
SearchStorage.com: NetApp boasts of a potency with VMware. Now we have vendors building
machine storage systems from a belligerent up. Are they a hazard to NetApp?
Hitz: Around 2007, we done a large gamble on VMware. We restructured a association around it.
We’ve been doing VMware for a prolonged time now. We’ve partnered effectively with VMware, some-more so than
EMC’s possess storage people. we don’t know what new guys are doing that’s designed from a belligerent up,
but we have years of sappy in that space.
SearchStorage.com: NetApp was a initial to offer deduplication for primary data. Are many of
your business regulating it yet?
Hitz: We took a demeanour during a adoption rate of dedupe around a world. The worldwide
average invasion of dedupe is about 25%, nonetheless by geographies, some are enormously higher. In
Denmark, around 65% of all NetApp ability commissioned was deduped. Nebraska was 2%. we started
thinking, ‘What’s wrong with Nebraska or special about Denmark?’ What happens is these technologies
go viral in an area, and afterwards they spread. we nonetheless 25% worldwide, maybe that’s best we can do.
But if a 65% in Denmark, because not other places? It incited out, a channel partners in a Nordic
region only got a sacrament on dedupe and done it standard. They pronounced ‘We spin it on by default;
if business insist they don’t wish it we spin it off.’ But customarily they don’t ask to spin it
If we unequivocally wish business to take advantage, a idea needs to be to make them buy it on
default. That’s where we’re headed for dedupe and compression. Software has to be intelligent adequate to
look during a use box and have some-more or reduction assertive dedupe and a complement can auto-tune.
SearchStorage.com: What is a NetApp
strategy toward cloud storage?
Hitz: People use a tenure ‘cloud’ in lots of ways, infrequently people use cloud as they
have a garland of VMware. A lot of people meant a large-scale common infrastructure information center. The
other definition of cloud is as a business model. Do we wish to build a new information core myself or let
somebody else build a information core and I’ll lease it from them? The doubt is, ‘Who’s going to
build a information center: a patron or a cloud provider?’ We concentration a lot on cloud providers as a
When we demeanour during people regulating 50 petabytes-plus, many are doing some character of cloud computing
either as an infrastructure play or a some-more targeted approach of building out an inner service.