The value of Blogging – Feedback
by Hu Yoshida on Dec 21, 2005
When I opened up my Feedreader today, I found that I had just been given a new distinction. I was voted the “Worst Corporate Blogger” by Anil Gupta on Network Storage. I was given this vote based on my blog on Where Should intelligence reside? I was trying to make the point that intelligence needs to reside everywhere. But intelligence is only as good as the information it has. So the question is not where intelligence should reside but where does the information reside to support a particular function, like storage virtualization. Anil thought this was a “blatent push for Hitachi products.” His main criticism was that I don’t really understand blogs. However, he was helpful in pointing me to Robert Scoble’s blog as an example of good corporate blogging.
Robert Scoble is a certainly a great blogger who can inform and elicit dialogue. Since I am relatively new to blogging, I appreciate suggestions like this to improve my blogging techniques. While I am in this learning mode, I ask your indulgence to look past the technique and let me know what you really think about my conclusion that storage virtualization should reside in the storage controller and not in the SAN. I am sure that there are many who read this blog who disagree. I promise to post any comments favorable or unfavorable to this point of view.
Let me know what you think
Comments (8 )
Thanks for your comment and reading my blog. I had quite a few issues including your comments to cover. So, I wrote them as a blog entry “Good Guys can be bad, sometime!” at http://andirog.blogspot.com/2005/12/good-guys-can-be-bad-sometime.html
Looking forward to further industry insight from you.
HDS is quite close company in my opinion and here is huge lack of publicly available information. No discussion forums and mailing list regarding HDS products. And no any community around your company. That’s why this blog is very important for everybody who uses HDS products (at least for me). It makes your company and products more transparent and attractive.
USP/NSC are one-year-old news (everybody waited for solution like NSC since EMC offered modular Symmetrix)… I suppose market share’s growth of HDS proved your strategic vision and there is no reason anymore to discuss worth of storage based virtualization. I would like to read your opinion about current changes and future steps.
Just an example… HDS told about ability to acquire AppIQ and wishes that this company staid independent. Now HP acquired AppIQ… Could you please comment this situation and future of storage software development?
Your blog on where should intelligence reside and the ideas of server, NAS, SAN and storage virtualisation were highly interesting and, to my mind, a perect example of web logging (I hate the term ‘blog’). I’m not sure about the SAN virtualisation vs Storage controller virtualisation distinction and need to think more about it. Congrats on a powerful presentation of your ideas.
Your recent blog provided me with releif that companies are actually progressing storage technology rather than building faster silos. A blog was never meant to be an unbiased source, more like a sharing of thoughts and ideas. I beleive your theory is very much a good one. This theory is in line with what HDS is doing and telling us is part of your job. I look forward to reading more of your thoughts on storage technology.
Now for the technical talk. I agree that if you are going to take the path of storage virtualization the virtulization must be in the target. Can you tell me how this is different from IBM’s SVC?
Additionally, If you simply put everything through a single storage controller I would forsee a few problems:
#1 Guaranteed Performance: If all of your Storage traffic is going through one controller, then a bottlneck would certainly ensue. Are there any options to grow the throughput and are there any SPC calculations for this specifically? Can I allocate throughput to other arrays over yours?
#2 Vendor lock-in: If I chose to use HDS as my virtualization vendor today, how would I migrate to your competitor in the future?
#3 Availability: What happens when your controller goes down or requires maintenance? What am I assured that my other arrays will work with your controller?
Again, thanks for your insights and keep up the great work.
Hey, don’t worry about what other people say, especially when they use epithets like “best” or “worst”. Just keep blogging, and keep talking about the stuff that interests you and that keeps you going.
Keep it up!
Those are comments really puts things into perspective –you’re right it’s the long term relationships that matters.
And keep up the good work at Technorati.
My thanks to all who read this post and sent comments. I appreciate a company like HDS which is owned by a Japanese company may seem secretive, top down, and closed, So my mission is to help folks understand us, HDS, as well as our products. As Dave and Jo have have suggested, this needs to be a long term relationship.
In response to Nathan’s Questions:
1)Guarranted Performance: The storage controller in this case is a multi processor with 128 processors accessing a global cache over a cross bar switch. So it is does not create the bottleneck associated with a single controller or cluster of controllers. It also has the capability to partition resources to provide QoS for different application. We do not publish our own SPC numbers since we OEM this product.
2)Vendor Lockin: External storage is not restricted to HDS storage. It can be FC attached IBM, EMC, HP, SUN storage. The data resides on the external storage in LUNs that the external storage defines. Therefore the external storage can be disconnected from the HDS USP/NSC and used natively.
3)Availability, the multi processor architecture with cross bar switch, enables concurrent maintenenace. One or more processors can fail and the system will continue to run. The cross bar switch will switch around component failures and hotspots. It is unlikely that the entire system will go down. In this eventuality we provide remote replication. In theory any array that supports FC attach, can be attached to the USP/NSC. We do qaulifications and have an approved list of other vendor products that we support. Not all products have gone through qualification by HDS. Please check our website for qualified products http://www.hds.com/products_services/universal_storage_platform/specifications/externalstorage.html
Anil Gupta needs to grow up.
Your association with HDS is honest/open, so you are entitled to toot your HDS horn on your blog as much as you want.
You are an HDS principal using HDS information resources (and, presumably, HDS work time). If you weren’t actively using that to promote HDS offerings and business value, you would be rightfully fired.