Quick Follow-On to Aperi Snub
In my previous post about Aperi, I quoted email from a PR agent for Symantec, which I classified as pretty hostile toward Aperi. Before I forget about it, I wanted to publish an email note received from that person in response.
We want to clarify that Symantec’s position isn’t rhetoric, and it isn’t meant to be hostile. It just happens to be unsupportive of Aperi. We have thoughtfully and carefully considered this view. If you’re interested in understanding why we maintain that perspective, then we will gladly discuss. (But I want to stress that we’re not keen to go down a rathole, here.)
For the record, rhetoric has several meanings and was not intended as a slur. For the record,
1. persuasive speech or writing: speech or writing that communicates its point persuasively
2. pretentious words: complex or elaborate language that only succeeds in sounding pretentious
3. empty talk: fine-sounding but insincere or empty language
4. skill with language: the ability to use language effectively, especially to persuade or influence people
5. study of writing or speaking effectively: the study of methods employed to write or speak effectively and persuasively
My use was associated with definitions 1 or 4 in this case.
I will be following up with an interview when I am back from Australia.
August 5th, 2006 at 6:32 pm
Techies use the word “rhetoric” synonymous with “hot air”, “bs”, and “empty, meaningless, you are wasting my time” dialogue.
Based on my experience, once the word “rhetoric” comes into a conversation with a “techie” or “techies”, that conversation is one-sided from that point forward and therefore ended.
(”lost to meaningful dialogue” sounds a whole hell of a lot more meaningful but is pure rhetoric to a “techie”).
To a “techie” the word rhetoric is a “show-stopper”. I once thought it was a request for either clarification or further information.
I now take the word “rhetoric” as a clear warning I am in trouble in that conversation. And probably all others with that individual.
If I am speaking to a group I have to make the decision as to whether the “rhetoric” speaker speaks for the entire group, a majority of the group, or key individuals in the group. In any case it is trouble unless you can disarm the power of that word by shifting your next words to an accommodating, conciliatory, agreeable or some other personally pleasing context.
This entire comment is rhetoric.
Does that mean it is 1, 2, 3, 4, and 5?
August 5th, 2006 at 6:52 pm
The quote “We want to clarify that Symantec’s position isn’t rhetoric” signifies to me that the Symantec comment is desired to have an intimidating effect on the reader. While it may not be “hot air”, “bs”, and “empty, meaningless, you are wasting my time” dialogue it could well be even worse.
Worse is meaningless posturing, gesticulating and other forms of intimidation. “We mean business here!”.
You won’t know until the first “action behavior” occurs.
This once meant striking the first blow or firing the first shot even if it is just a warning shot across your bow.
The British put on their bright red coats, marched out into the middle of open ground. and stood closely together to concentrate their superior fire power.
The Colonials wore clothing that blended with the vegetation, hid behind trees, rocks and bumps in the ground and fired “at will” from their hiding places just out of range of the shorter muskets of the British. We won that war and introduced guerrilla warfare to the new USA.
We need some of that mentality in the Storage Revolution.
August 7th, 2006 at 11:14 am
There’s an old joke about how the brave British wore red to hide the color of blood. Similar reasoning went into the French’s decision to wear brown pants.
Sorry. I couldn’t resist.
August 8th, 2006 at 2:10 am
Ha! Ha! Ha! That’s funny.
I’ve worn those both ways.
Depends on which Storage vendor I am talking with.
What would you wear if you were asking a Storage vendor about why they don’t do Content typing of Information in cache?
Or somewhere in the Storage Unit of Technology?
How about replacing all RAID with JBOD?
That would simplify all the discussions about which RAID is best.
How about “Data Pump” (my name) Storage Units of Technology?
The IDC uses a thing called a “Chunk Server”. Some people call them “resolvers” because they resolve Information into some scheme so it can be managed.
I’ve had some strange responses to these questions.
Maybe I’m not an EDC guy. Just IDC.
August 8th, 2006 at 8:07 am
Hi, Robert. Concerning “Content Typing of Information in Cache”, I assume you mean cracking the file to automatically generate metadata. This includes the simple stuff like file name, extension, location, age, size, etc., but also more complicated stuff like Microsoft Word metadata, jpg metadata, individual mails within a PST file, DICOM medical information, etc.
If so, I hope you’re talking about file-based storage, not block, right? It’s pretty scary for a block-based product to start snooping on the filesystem directories, FATs, inodes, whatever, to figure out what a file is. And then there’s the issue of being aware of file locks, which aren’t visible at the block level.
So assuming you’re talking about file-based storage, like a NAS boxe, then I mostly agree. Low-end products should be self-contained and support this metadata generation, along with appropriate tools to try applying HLM strategies.
But I think that many of the folks that need Content Typing are those with LOTs of clients and LOTs of storage destinations. In that case it makes more sense to have a centralized policy server that is able to push out the policies to the clients where the data is being generated. That’s where files are understood, and where multiple paths to storage originate. In this configuration it wouldn’t make much sense to push the metadata to the storage. The storage should just be simple, fast and redundant.
This is probably too much to talk about in a comment section of a blog. Maybe we should stick to making fun of the French.
August 8th, 2006 at 8:14 am
Regarding replacing RAID with JBOD:
I assume you mean moving RAID to other parts of the system, and not necessarily leave it in the storage? If so I agree. I doubt you’re talking about getting rid of RAID, i.e., storage redundancy. That’s crazy talk.
The reason I jumped on this one is that there are some morons out there that babble about RAID being dead, and they’ve got a new technology to replace RAID. But when you look into it, all they’re saying is that they’re replacing RAID-5 in the storage with RAID-1 somewhere else in the system. That’s not getting rid of RAID! It’s changing what level is used and where it’s used. And RAID-1 is great if you can afford the cost of the extra drives.
OK, I better leave it at that before I come completely unhinged. I drink a LOT of coffee.
August 8th, 2006 at 5:10 pm
Concerning “Content Typing of Information in Cache”
Thanks for the excellent reply. That’s 100% more than I got from the NetApps guys. Or any of the other major Storage vendors I asked. Although, I have to admit, some of them did look very wise while not replying.
With regard to “This is probably too much to talk about in a comment section of a blog. Maybe we should stick to making fun of the French.”.
I’m not sure what to do about an ongoing, productive dialogue. This is Jon’s Blog and I’m sure he will take charge of it when we are out of hand. Many Bloggers seem to feel comments should be short. I’m sure that long comments are not read much. On the other hand I sort of like the idea of a “Blook”.
(http://en.wikipedia.org/wiki/Blook)
From Jeff Jarvis at the BuzzMachine Blog.
[Begin long URL]
http://www.buzzmachine.com/index.php/2006/08/04/exploding-books-i-everybodys-an-
author/
[End long URL]
“Next up is the blog slurper (with other slurpers after that). It will take your blog, grab the content, and let you edit it, publishing it as is (see Tony Pierce’s blook). I can also see using the blog as a writing and publishing tool for the express purpose of ending up with a book (something I’m thinking about doing with a book on books). And of course, see Tom Evslin’s Hackoff.com, written as both a blog and a book.”.
http://en.wikipedia.org/wiki/Tony_Pierce
http://www.hackoff.com/
I believe Storage could benefit from some Information vision sharing. Information is all that matters. Storage is just the “Enabling Unit of Technology”.
Think of a car without a person to drive it. It could be a Yugo or a Mercedes. Without a person “Unit of Information” the car is just dead metal.
August 10th, 2006 at 11:39 am
Can anyone speculate why Robert’s posts are double posting?
August 10th, 2006 at 11:42 am
Margaret tells me that sometimes, when the server is getting hammered by spam, response time is slow. That might encourage someone to hit submit twice. Please don’t.
August 10th, 2006 at 12:20 pm
Yeah, I spoke to Robert on the side and his double post was due to response time.
And, YES!, you’re getting a ton of spam. I have an RSS feed where I can see you getting ~20 a day. That’s still better than my blog where I get ~150 a day.
August 10th, 2006 at 1:28 pm
Sorry for the double post.
The feedback on the problem is great.
I just assumed I failed to hit the submit button, as I have done sometimes in the past.
I promise to wait longer, 24 hours or more, before re-posting.
August 11th, 2006 at 2:17 pm
It is easy these days to pick up four 500GB external firewire/USB drives from Costco on the way home, plug them into a 14$ four port firewire card, and end up with approximately 2TB of space on your home media server.
But we’re still dealing with files and filesystems as though the year was 1985. Seems to me that the time to determine what a user wants to do with a file is when that file is being added to the filesystem. Not by some usage based heuristic. I see enriching the semantic content of filesystems as a good thing.
And backups are just a farce. Few people would be willing to use a computer that randomly lost bits out of main memory. Any time the processor is in the idle loop, there is time to be making sure that data isn’t stored in only one location.
August 13th, 2006 at 5:13 am
What Robert Clark says is just dynamite!!!
I would love to see a dialogue on this topic.
It is multi-level.
The requirements and the Strategy are different at each level from
Personal Computing, SOHO, SMB, Enterprise, Enterprise Data Center, and Internet Data Center.
This also ties in with Hu Yoshida’s comment,
“I had another great conversation with Jon Toigo last Friday about Records Management and active archives. Jon, who is always quick to spot new trends, described RIM, Records and Information Managers, as the new rock stars for IT.”
August 14th, 2006 at 2:01 pm
OS X Leopard will likely feature Time Machine. TM exposes a developer API which allows coders to provide hints as to what does not need to be backed up.
I haven’t followed DragonFly closely enough lately to know if they are making any progress in porting ZFS. The prequisite vnode work does appear to be getting done, so who knows.
People closest to most of the problem data are the end users. End users don’t make a distinction between PC, SOHO, SMB, etc. Those are marketing distinctions? (If there were enough large customers, no one would have bothered making the distinction?)
End users, for the largest part, bring a DOS and or Windows level concept to them to the office. Aside from system recovery point, or the filesystem that lives under a VM, or third party HSM, what advancements are there?