|
Lightning in a Bottle or Chicken Little?
HDS Upgrades 9900 V Series
By Charles King
Hitachi Data Systems (HDS) has announced upgraded
capacity, connectivity, and throughput options for the company’s Freedom
Storage Lightning 9900 V Series storage systems. According to HDS, these
enhancements combined with the second generation crossbar switch-based
architecture and virtual storage ports will allow customers to lower TCO via
massive storage system consolidation. Enhancements to the Lightning 9900V
Series include 146GB drive options, which effectively double the machine’s
raw capacity to 148TB, and an increase to thirty-two FICON channels and sixty-four
Fibre Channel connections at 2GB per second. In addition, the 9900 V Series
now features Virtual Storage Ports (VSPs) for SAN
environments that allow up to 128 different server platforms or operating
systems to share the same physical 2GB/s Fibre Channel port. In addition, HDS
promoted a “networked storage blueprint” which leverages the company’s
Lightning 9900V Series and the Thunder 9500 V Series for distributing storage
across enterprises. No pricing information was included in the announcement.
Analyzing IT press releases often resembles
divination by examining chicken guts. Sure, you might find something
interesting in there, but you still usually end up with a bloody mess. The
process can be complicated by peripherally linked events, though their
influence is often difficult to ascertain. The event we suspect has
influenced HDS’s latest rev of its Lightning 9900V
Series is EMC’s upcoming announcement concerning the company’s Symmetrix
product family. While it is too early to discuss point-by-point comparisons
between HDS and EMC offerings, we believe that what this week’s Hitachi
announcement contains and what it is missing both provide food for thought.
Raw capacity enhancements which essentially double the 75TB available in the
original 9900V may sound impressive, but the press release also stated that
the average size of 9980V installations has increased to just 20TB. In other words,
doubling raw capacity constitutes little more than a “bigger is better”
argument that will be of little practical use to the vast majority of HDS
customers. Additionally, while the announcement makes a great deal of hay
concerning HDS’s VSP capabilities, storage vendors
including EMC, StorageWorks, and others offer
similar multi-server and OS Fibre Channel support. Finally, HDS’s network storage blueprint looks remarkably similar
to distributed enterprise storage models offered by EMC, IBM, and other vendors.
What is missing from the release is also intriguing.
For example, despite doubling raw capacity, there was no mention of similar
chip/processor enhancements in the 9900V, an issue that could affect
performance scaling, especially in the highly consolidated storage
environments HDS is promoting. Additionally, we were curious to note the
absence of any improvement in the system’s cache size or backend capacity,
both of which could provide notable performance boosts. In general, the
enhancements HDS is trumpeting for the Lightning 9900V Series appear to be
old news or of interest to only a limited number of storage users. Overall,
the viscera of this particular chicken looks to be sad stuff, especially
compared to the Symmetrix barbecue EMC is rumored to be cooking up.
|
|
Slammer Déjà Vu
By Jim Balderston
The SQL Slammer worm made its way across the
Internet late last week causing substantial disruption to the Internet and
even reaching beyond to affect other networks. According to reports, the worm
began moving through Asia and then across Europe and into the U.S., creating
huge amounts of network traffic that resulted in data losses of up to 20% of
all packets sent. Email, Web serving, and according to number of reports, ATM
service was disrupted as a result of the worm’s ability to spread itself so
effectively. According to those reports, Bank of America, the nation’s third
largest bank, was unable to provide transaction service at nearly all of its
13,000 ATMs. The bank said it had returned the ATMs to service by the middle
of the weekend and that no customer data was affected. Continental Airlines
reported that computer systems in their various reservation centers were
affected as well, although for periods of time less than thirty minutes for
the most part. Security experts said the Slammer outbreak was the worst since
Code Red.
Every so often — in this case every year and a half —
events like Code Red or the Slammer worm remind the world at large that the
Internet is still a very fragile thing indeed. While many still point to its
original design — a distributed, multi-path structure as a defense against
nuclear attack — as a sign of its robustness, the Internet still remains
remarkably subject to serious degradations in service. In this case, a very
small piece of computer code so flooded the system with server requests that
one in five messages was lost.
As long as the Internet remains in its present form —
largely an unintelligent network for sending and receiving packets of
information — such problems will continue. The public nature of the medium
means wonderful discounts in the cost of transmitting information across vast
distances, but it also means that anyone or anything can get online,
including the person who released Slammer on the world. For enterprises, the
lesson is simple: you get what you pay for. If companies want to run their
data streams on public networks, they run the very real risk of having them
compromised by denial of service attacks like Slammer. They should think long
and hard about how they may want to segregate their information flows. Some,
the more critical, may need to be on dedicated lease lines that are free from
the ravages of the public domain. At the same time, careful consideration
must be given to the idea that the Internet as it is now composed needs to be
upgraded. The fact that thousands of IT personnel around the world had to put
their hands on servers and wrestle with Slammer tells us that more
intelligence in the network is well overdue.
|
|
IBM and Siebel Systems Unite on
WebSphere
By Myles Suer
IBM and Siebel Systems announced this week an
expansion to their existing business alliance. The new agreement moves Siebel
away from their proprietary application server to WebSphere. Siebel stated
the driving factor for extending its existing relationship with IBM was the
increasing importance of integrating its application with solutions from
other vendors through Web Services. The joint presentation by the companies
described two developments on Siebel’s part. The first was the increasingly
vertical nature of CRM products; the second was the admission that custom CRM
developments both dominate and effectively stifle the package CRM industry.
To grow its market, Siebel announced a blended CRM strategy that involves
making custom applications integrate with the Siebel offering. To this end,
Siebel will provide native support for J2EE and Web Services running on the
IBM WebSphere. The companies also indicated that Siebel would create tighter
integration between Siebel eBusiness Applications and IBM Lotus messaging,
calendaring, advanced collaboration, and eLearning technologies as well as
integration of high availability, Tivoli systems management, and security
software.
We find this move interesting on a number of
grounds. It illustrates the increasing importance of Web services and
proffers that this integration will not happen as many a venture capitalist
hoped for, in discrete Web services vendors, but rather that the application
server will be the connection engine for the enterprise. Given the
geopolitical nature of this alliance, it will be interesting to see if this
leads to similar announcements for other enterprise application vendors; SAP,
PeopleSoft, SAS, and BMC among others. The fact that Siebel announced a
relationship with Microsoft .NET in October would tend to temper such
thinking. However, we think the broad scope of their IBM relationship will
cause others to act. Equally as interesting is what is not included in this announcement. Someone has to build the
missing links between the various applications in the enterprise. It is all
fine and good to have an application and a platform that can speak Web
services but unless the custom in-house applications can be made to connect
over Web services, the undertaking is largely moot. What may prove
interesting to see is whether IBM having a significant systems integration
business steps up to the task of building custom Web services to connect
custom CRM and legacy applications. If this happens, then this IBM and Siebel
relationship would be well positioned to provide real and lasting leverage to
customers, while providing Siebel a significant competitive advantage.
|
|
IBM Provides Focus for Its Document
Management Pursuit
By Myles Suer
News reports this week noted IBM’s plans to boost
research and development spending on content management software by 25% in
the coming year. IBM will also shift 2,000 salespeople to supporting content
management sales. These moves seem to indicate realignment in how IBM views
the coalescing of the content management market where it positions
Notes/Domino as a collaboration/content creation platform, WebSphere for
application serving, and WebSphere Portal for content delivery.
Historically, document management has not been an
exciting enterprise software business. The problem is that managing documents
has been a “nice-to-have” feature in the enterprise, but it has not been
critical to anyone with the budget to buy such technology so it has been a
difficult proposition to sell. However, we believe the market has turned more
interesting in recent years. This change has resulted from the Web becoming a
very efficient means for companies to disseminate their latest (product)
content to prospects and partners and the desire to do the same internally.
Initiatives including records/email management and collaboration have also
helped to focus the content management story.
IBM’s move positions it to go after all elements of
enterprise content management. Beyond unstructured content, we expect IBM to
identify the opportunity to add additional software products to its product
portfolio. In particular, this includes the other enterprise information
portal (EIP) application that uses large databases/warehousing. We contemplate
IBM acquiring or building a Business Intelligence (BI) product because its
application server provides the connectivity needed to reach into the
enterprises’ structured content. Integrating collaboration, content
management, and business intelligence into a single platform would be very
compelling for customers. Such a move could further challenge Documentum and
Filenet. With IBM, Oracle, and Microsoft focusing increasing attention upon
this space, Documentum has grown its market share and differentiated itself
by aggressively pursuing vertically tailored products. We believe the unique
needs of verticals may be more significant in the content management than any
other space in the enterprise software market. It will be interesting to see
if the vertical strategy remains unique to Documentum and whether it or IBM
will be the first to create a complete market offering by opening their arms
to business intelligence independently.
|