Continuing our series of Industry Interviews, we spoke with Charles Milligan,
Manager of Advanced Technology for StorageTek. Charles is responsible for
identifying and proving emerging technologies that have potential to disrupt the
storage systems markets, by architecting solutions to adopt them and creating
intellectual property to control them. In simple terms this means that his job is
to figure out what’s coming down the line in the next 3-5 years, and understand
how these developments will affect the storage landscape. Charles has
authored dozens of white papers on storage technologies and holds 37
storage related patents.
ESF: Hello Charles. Thanks for taking the time to talk with us today.
Charles Milligan: You are welcome.
ESF: Your role at StorageTek means that you are responsible technologies
that will affect the storage industry in the future. On a corporate level, what are
the main challenges facing storage users over the next five years?
Charles Milligan: Storage will continue to get less expensive and larger
capacity and we will continue to fill up what we have. The real problem is what
we are able to do with the vast amounts of data we will be acquiring. The
challenge is to make the data more useful rather than just have more data. To
do this we must solve some fundamental problems in how we think about
information and knowledge. The information and knowledge that can be
derived (created?) from data is not a static attribute, it changes over time as
other aspects of the environment change. How can we maximize it?
That problem is exacerbated by the increasing mobility of most decision
makers. The information and knowledge must be available for use in
distributed areas that are remote from where it was created or processed.
Furthermore, the areas will change constantly (much like the difference
between using a desk phone, and having a cell phone).
ESF: One of the most significant technologies coming through is iSCSI. As
iSCSI products start to appear on the market, what effect do you think they will
have on storage technologies and implementations? Who has the most to gain
by the low cost iSCSI hardware?
Charles Milligan: The real value in an iSCSI storage interface is that the
enterprise operations can be streamlined with respect to system
administration. The best place to use iSCSI hardware is at the user (host)
interface to the storage subsystem. Within a subsystem, other technologies
can be used with greater efficiency (e.g., Infiniband as an internal networking
technology). Low cost iSCSI developers should focus on improving
performance by moving the TCP/IP processing into hardware.
ESF: What are the next round of technologies coming to the storage market?
How will they affect the way we store and manage data?
Charles Milligan: Miniaturizing is the current trend. MEMS, Probe,
nanotechnologies are getting a great deal of attention. IBM sold off their disk
operations and is focusing on new technologies such as millipede, which take
advantage of nanotechnology. The interesting thing is that the next round of
technologies that disrupt the enterprise storage markets will probably come
from the consumer sector. The Xbox that Microsoft is pushing for video gaming
could grow up to be the next server platform. It will be interesting to see if we
can incorporate processing power into the miniature packages that are
developed for the MEMS and nano storage technologies and enable a new
paradigm for information management.
ESF: In storage we always seem to think that more is better, but with
capacities increasing exponentially, is there not a chance that demand for
storage will be outstripped by supply?
Charles Milligan: We already see this in the PC market place. Seagate has
informed the industry to expect the rate of capacity improvements in their disk
offerings to slow down dramatically over the next few years. On the other hand,
the use of storage will change markedly as the cost decreases. Very few
individuals have all their photo, video, music, etc. stored in a managed digital
fashion. When storage is essentially free, the norm will be to have a few
terabytes at home and in the car, and at the office, and online vial some
wireless connection. I think that oversupply will only be a problem if the size
grows but the cost does not continue to decrease faster.
ESF: For many organizations one of the biggest issues surrounding storage is
not how to store data, but how to back that data up and satisfy DR and
business continuance requirements. What technologies, products or
approaches are coming that will help companies to manage this aspect of their
Charles Milligan : The primary change will be to the 40 year old paradigm of
saving a copy every once in a while and putting that away for recovery from
future disasters. The new paradigm recognizes that one does not want to back
up data, or for that matter, even to restore data. One really wants to simply
ALWAYS be able to get at data. The technologies that are able to maintain a
stream of consciousness with respect to data such as change logging coupled
with snapshots will soon change how “recovery” is viewed. When I have lost a
piece of data (a record, a file, a volume, etc.) it is much more convenient for me
to be able to simply look into a historical library of what was and pull something
from the past into the present. There are several companies working on
offerings with these capabilities (e.g., the StorageTek Echoview).
ESF: One of the hottest topics in storage today is that of security. How do you
see security being implemented into storage products, or do you think that the
task of securing data is best left to software products like operating systems?
Charles Milligan: The security threat model includes at least seven different forms of
attack or theft or denial of access to data. There is no one form of “security” that
will protect against all of these threats. When encryption is an appropriate form
of security, it is most effective when it is applied as close to the creation of the
data as possible. That puts it in the host or server rather than in the devices.
When authentication is required, the devices must be precluded from being
able being able to abrogate it. However, insuring that data has not been
tampered with since it was stored (bonded data is one term people use) is a
service that is best implemented in a subsystem. There are mechanisms that
allow the data to be assured even when it is copied to new media. When we
have active devices that can run applications where the data resides (rather
than moving the data to the application), the device will have to incorporate the
authorization controls associated with security.
ESF: There is a great deal of talk about vendor compatibility between storage
products. Do you think it’s as big an issue as everyone makes it out to be?
Charles Milligan: Compatibility is important when there is a particular
technology that becomes a commodity. It is very irritating to buy a (FC, Ethernet,
etc.) network card for a server that is not “compatible” with the device attached
to the server. However proprietary systems have rarely been able to be
interoperable among competing vendors and this is not necessarily a bad
thing for the customer. The competition is what continues to drive innovation.
ESF: If you had to pick one technology that you think will have the greatest
impact on the storage industry over the next five years, what would it be?
Charles Milligan: Miniaturization and Commoditization – all the things we have
used as leverage to keep margins high are becoming commodities that
anyone can use to whip up a new product offering. Companies become very
agile, and products become interoperable. Cameras and cell phones will drive
memory on a key chain style storage that will be impossible for an enterprise to
ESF: Thanks for your time Charles. It’s been great to talk with you.
Charles Milligan: Thank you.