After - 4 New Storage Interfaces - How did they Fare?|
State of the
storage market (summer 2006)
STORAGEsearch.com (June 2006)
STORAGEsearch.com launched 4 new
directory pages dedicated to emerging storage interface technologies which had
got a lot of reader attention in our
news pages at that time:-
These new storage interfaces, conceived at the start of the
2001-2003 US IT market recession did not all match up to the original hopes of
their proponents. |
In 2006 we went back to see how successful they'd
been in the market compared to their original promise.
has been the most successful.
It has been a multi billion dollar
market for several years. In 2006 over 300 million
hard disk drives will
have SATA interfaces. SATA has also made its appearance in
solid state disks,
DVD drives and
had an optimistic, much hyped but false start.
The market in 2006 is
2 to 3 years behind the market size which was originally predicted for
In the early days of the iSCSI market optimistic projections
IDC led vendors to
expect that iSCSI would be a $1 billion market in 2004. Those early projections
for 2004 iSCSI revenue were ten times too high. We may have to wait till
2007 before the iSCSI market reaches the magic billion dollar size. The main
reasons for iSCSI's slow take-off were:- slowness in the standard development
process, early products which didn't work properly, and then finally, waiting
for Microsoft to wake up and grasp the significance of the storage market.
Looking to the future, the market is now hotting up. Recent iSCSI benchmarks
on 10Gbps Ethernet have quoted faster IOPs than
SANs. Vendors are promising
that iSCSI will be faster than FC without the setup complexity.
Attached SCSI - It took 4 years for SAS products and systems to start
...Later:- in September 2006 - a survey of European VARs by
LSI Logic revealed
that only 14% had already sold SAS solutions to end users.
Hitachi shipped the
first 15K RPM SAS
hard drives a year
ago (May 2005) and most leading server manufacturers had launched SAS based
servers by the end of 2005. But I get the impression that most users are
underwhelmed by the performance promise that SAS based systems offer. The main
reason is that the fastest
SATA disks overlap in
performance with mid range SAS drives. So the benefits from adopting SAS only
appeal to a small segment of the market.
Another reason for the small
size of the SAS market may be that
parallel SCSI had become a
niche for RISC based servers running various flavors of Unix. So although users
of Sun's SPARC servers, for
example, have seen an upward shift in performance from the new SAS drives, the
Unix part of the server market has been overtaken by Windows based servers using
Intel Architecture processors. Most I.A. server makers were natural early
adopters of the Intel rooted SATA interface. In contrast, most RISC servers
never supported PATA, and those product lines had to wait a few more years for
the software compatible Serial SCSI disks to emerge.
has been a graveyard
for many startups
which came into being to support this technology. Of the 4 interface standards
which started at about the same time the InfiniBand market is today the smallest
in revenue and in growth potential.
...Later:- in September 2006 - the
Trade Association estimated that only "over 500 end-user sites"
had deployed InfiniBand products in production applications.
original idea behind InfiniBand was that it would offer an industry standard
alternative to the many high speed proprietary busses which server manufacturers
used to cluster their most powerful servers. The server recession in
2001-2003 slowed down the pace of new server developments and provided a
disincentive for manufacturers to end of life their most profitable products. In
the past 4 years two other factors have reduced the potential market size for
The availability of processor chips with multiple central
processing units on the same chip has reduced the need for motherboard to
motherboard memory access of the type provided by a factor of two, four or
eight - for different chip implementations. Furthermore the availability of
10Gbps Ethernet, and the possibility of 20GbE provides a workable alternative in
many applications which would have looked like natural slots for InfiniBand
just a few years ago.
Although InfiniBand has reached the stage where
there are many working products - the shape of the market is a small number of
end-users who consume a large number (thousands each) of InfiniBand ports. It is
doubtful that this will become a high volume market unless something radically
changes in the architecture of server systems. If that change happens - you'll
be the first to know as it gets reported in these news pages.
Summary - The storage market has changed a lot since these new
storage interfaces were hatched in 2001. The new interfaces have survived a
recession and it's likely that they will all still be around in some upgraded
form till the end of the decade. They were designed to meet the needs of
computer networks in which the main form of online storage is hard disk based.
In the next few years we'll start to see a shift away from rotating mechanical
storage (hard disks, DVDs and tapes) towards faster, more reliable solid state
Solid state disks currently
use interfaces which are hand-me-downs from the magnetic storage era. The next
generation of new storage interfaces which may start life in 2011 will be
completely different. Watch this space.
You can get more detailed
information about all the above markets, vendors, history, applications etc by
clicking on the links above.
10 Years After -
(featuring Alvin Lee on lead guitar).
Attached SCSI - is it worth the wait? - article republished from InfoStore
STORAGEsearch.com we've had a
magazine page dedicated to
SAS since 2001.
SAS products have been shipping to users since 2005 - but the
market has been stuck in a prolonged "innovator" phase. As the SAS
market is poised to advance to the "early adopter"
phase - I thought it would be good to have an article (which is less technical
than those we've previously published) to introduce the benefits of this
technology to more readers. This article is republished here with the kind
permission of InfoStore - Asia's
leading storage magazine. ...read the article
|the Fastest Solid State
Speed isn't everything, and it comes at a price.
| But if
you do need the speediest
SSD then wading through the web sites of over 74 current
SSD oems to find a suitable
candidate slows you down.
And the SSD search problem will get even
predict there will be over 100 SSD oems in 2008. |
I've done the
research for you to save you time. And this page is updated daily from
storage news and direct
inputs from oems. ...read
|Are MLC SSDs Ever
Safe in Enterprise Apps?|
| This is a follow up
article (published in March 2008) to the popular
SSD Myths and
Legends which, a year earlier demolished the myth that flash memory
wear-out (a comfort blanket beloved by many
RAM SSD makers)
precluded the use of flash in heavy duty datacenters.|
article looks at the risks posed by MLC Nand Flash SSDs which have recently
hatched from their breeeding ground as chip modules in cellphones and morphed
hard disk form
|| It starts down a familiar
lane but an unexpected technology twist (which arrived in my email while
writing this article) takes you to a startling new world of possibilities.
Infiniband Established Itself in the Market? - article by Engenio|
article looks at the state of the Infiniband market at the end of 2005.
5 years stirring in the emerging market cauldron the Infiniband market hasn't
turned out to be the popular flavor which was originally anticipated. But
it's finally starting to get served up in some important markets.
Infiniband port now costs half as much as a fibre-channel port and delivers
many times the performance rate. According to the author, Infiniband is now
ready to take its place on the mainstream technology menu. ...read the article ,