Posts by Ant Evans 130 publicly visible posts • joined 22 Oct 2007
I thought 21st century advertising happened through media buying. In the unlikely event Twitter is trying to do this retail, then they deserve what's coming to them. Media buying might itself be labour intensive, but that's not necessarily Twitter's direct problem.
My very brief experience of Twitter is that it's labour intensive to use. Why would I spend all day sorting through a garbage tip, when I can just buy food in a shop?
If anyone is qualified to oversee the acquisition of Canada, it's Carly. Canada has a great brand name (although you wouldn't necessarily have to use it!), some great products, and an educated and capable work force (regrettably not a good enough reason to retain them all post merger, but needs must). There are clear potential synergies, in the jargon, for example, who needs Montana AND Alberta in the same country? I'm not even sure you need New England and Ontario. Redundant areas could be sold to an independent Quebec, killing two albatrosses with one stone, or in M&A doublespeak, releasing value to shareholders.
Carly for President. Everyone's an 'American' now.
I'm not sure how you can have this conversation without discussing what goes by the name of data policy (gets one tangential mention in the article) but may more precisely be described as looking after your stuff properly. Looking after your stuff properly has dimensions - sensitivity, customer, jurisdiction, access and retention spring to mind. I'm sure we could think of five others.
So you want to be asking yourself a few questions: do I know how to look after my stuff properly? Do I look after my stuff properly today? If all my sysadmins died or (heaven forbid) were inexplicably made redundant, could a random cloud person (possibly in a different cloud) 1. reconstruct my data and 2. look after it properly? Take a quiet moment to think about what it would mean to have perfectly safe data, but no idea what it was, whom it belonged to, who was allowed to see it, or how long to keep it.
This is not just your cloud provider's responsibility. If you're shit on the ground and go to the cloud, you will get shit rain, at best.
@Guus, flash is indeed used to mask poor design, just as fast processors and more RAM are used to mask poor design, especially in the database world. But the poor design is not always in the application layer. OSes have got bigger than most apps, and they are storage pigs with big peak loads, bigger management overhead, and a lot of redundant writes. Customers can't really optimize OS workloads significantly. Virtualizing whole machines exacerbates this linearly. So you can thank MSFT and VMW for some growing fraction of storage demand, and for a chunk of the demand for flash, which handles stormy workoads better than Winchester disks.
Flash is not the only way to deal with boot storms - you can create virtual clones and serve common blocks from RAM, maybe until Patch Tuesday. But flash is both easier than cloning, and reassuringly expensive.
Interesting that Chris only mentions flash cache once (flash array type 2). This is not what most people think of as a flash array, but any serious array has had flash and NVRAM for years. For dull workloads, cache works, and you don't need solid state storage at all. Nobody seems to want to hear this. It's just not flashy enough.
Finally, while consolidation is actual magic, it has some obvious and undesirable side effects: you get an increasingly entropic workload. You might not even know what's running on your array. Maybe caching will work next week, and maybe not. Why hang your arse in the wind? Faster gear makes for shinier underpants. Always has.
Spectacular and imminent destruction of wealth through ill-advised mergers. You have been warned.
The dynamic is similar to IPOs. Sweet-talking matchmakers bustle selflessly around the shy and vulnerable (IPO) or fat leftover (M&A) people, create a match made in heaven, and then scarper with their 10%. Relationship counselling not included. No refunds. Now fuck off.
Drive imaging is in some ways an answer to a question that needn't exist any more. As a few others have said, you need to distinguish between recovery and backup, that is, between system and data. Putting them both on the same volume/partition is not something a sophisticated user wants to do.
Separated, they present different problems.
Your options are limited on a factory built laptop, but you can create a new partition after shrinking the existing one. Under Windows, you may need to delete the swap file and System Restore data first.
Now back up your data. Robocopy and rsync are quite cheap, but it's 2014 so I'd add something like Carbonite. After nearly ten years I still don't know of anything offering better value for a single workstation. For hundreds of thousands of files, it's a bit of a hog. There's a small business version; never tried it.
As for the OS, consider concentrating on being able to rebuild it, not necessarily restore it. Highly customized OS instances make less and less sense the more of them you regularly use. Instead they should be generic and dispensable. To that end, where feasible, I install only VirtualBox on the metal and use VMs day to day. The underlying OS is less likely to get mangled, and if a VM gets mangled, roll back or restore.
So:
On the metal: bare OS, backup solutions, hypervisor. System recovery, plus install disks (and licences where applicable).
On a fast partition: VMs. Tweaks possible depending on architecture.
In your VMs: user customizations and apps. Local disk to disk backup.
On a slow partition: user data. Disk to disk plus cloud backup.
To understand IPOs, watch a wildlife film. The most obvious example will be on your local savannah, but there are also good examples in aquatic ecosystems.
How do predators take prey from larger species? Do they gather in packs, isolate individuals from the herd, try to outrun and exhaust their prey, attack in unison, and weaken the victim by inflicting small injuries? Yes, they do all of these things. But their most successful strategy is to hunt and kill the young. They're slow, weak, dumb, and very, very tasty.
What I know about networking is dangerous, so can someone explain this to me: how does SDN help you if what you're doing in software doesn't match your topology? And if you have to design, maintain and change your topology, which you do, isn't what you have hardware defined networking?
From my position of sublime ignorance, it looks like for SDN to be transparent you'd have to wire everything to everything else at backplane speeds. Even I can see that that's a lot of wire.
Someone help me out here.
“The world has a big issue around vectorisation and parallelisation of code,” Graylish said. “99% of code isn't written that way.” Graylish also feels “defining a workload that can run in 1000 cores is hard.”
With the greatest respect to Intel and Graylish, this conflates processes and threads. Since it's coming from Intel, whose business it is to know this stuff, I can only assume it's a deliberate obfuscation.
I guess you need to figure out if you're running a server or a portable lab. Laptops are relatively bad at always-on operation.
I did try running a (non-cheap) laptop for a year or so under load to see if the fan would fail, and the fan failed. On an idle laptop on the other hand, the fan will continually cycle on and off, increasing heat stress. You'd better make sure the fan is accessible, so you can clean it. Laptop hard disk cooling is usually via hope.
Li-Ion batteries are the worst possible choice for UPS. The warmer they are, and longer they spend fully charged, the faster they lose the will to live. If there is an option not to charge fully, use it.
Not a lab, but I virtualized my entire study and got a physical room back. The study now runs in various boxes hidden behind or in cupboards in different rooms, and I remote in from wherever I happen to be. Virtualbox makes it embarrassingly easy to run headless VMs for access over RDP, OS agnostic. When an RDP session gets confused, there's TeamViewer. For the server I use Win8.0 on a headless mini ITX AMD setup - AMD has excellent virtualization on all its chips, not just some of them. The power draw is negligible (get WD Black, and an AMD 'T' chip). RAM is better value than SSD: squeeze the VMs and cache their I/O at the host. By siting the server right, you can save on a media server. You can also save on fat clients - my main work machine is now a netbook, which is cool enough to run in the printer cupboard, pointing to a VM.
Another ex loyal Opera user here. I used it because it had some security through obscurity value, and consistently the best interface. Compatibility issues gradually vanished as sites dropped IE 6. The list of useful features it had first is long. Plus, it had a killer feature: when you click Back in Opera, it goes back; it doesn't try to reload the frikkin internet.
With a change of engine, Opera became just another browser, at which point I may as well use Firefox.
I still use Mini, because it is, like Opera, marginally superior.
Miscreants have brewed a multi-platform strain of malware capable of infecting Windows, Mac OS and Linux PCs.
The malware - dubbed Client Side Java by Kaspersky Lab - has reportedly infected 'three billion devices', which would make it the world's most successful bot net.
Command and control traffic has been traced to a firm in Silicon Valley.
Manufucturers have known for a couple of years that the generation X & Y increasingly see personal cars as a liability. At the same time they happen to be facing a set of problems exactly analogous to Big Tech: vastly improved capital efficiency has delivered market saturation, overcapacity, and redundant performance. Just as Intel has to convince us that we need a second supercomputer on our desks, BMW needs us to believe that we need 300bhp. Turns out we don't.
Car manufacturers' response has been 1. sell to the parents; 2. sell to emerging markets yet to understand the futility of universal car ownership; 3. deal reluctantly with the problem with better tech: baubles like ICE and GPS, inefficient hybrids, tech for OEM car sharing; 4. try to shut factories.
Big Auto doesn't have the answer yet, because a generation has to die for the new plants to grow, as it were. Some piquant lessons here for hi tech.
I don't care about TV.
All I know is we could do with better options for local networking in the real world. After a few weeks messing with dual band routers, I have successfully reinforced my prejudice that all wifi in the various low GHz frequencies is just different shades of shit.
I don't hold out a lot of hope. The physics tradeoffs are nasty, and the politics worse. Engineers have done a good job with garbage spectrum. I'd like to see what's possible with better quality raw materials.
Gates and Ballmer are both big shareholders. Whether they're on the board or not, they'll be billion dollar pains in the arse. In the event, it may be better to have them in the tent pissing out. But any new boss will need even bigger shareholders in their own corner, or they will get nothing done.
MSFT has so far failed Succession Planning 1. This is no way to run a public company.
Enterprise SATA drives are functionally identical to consumer drives. I hope they get a bit more testing to get rid of the most egregious duds, but that's it. They don't have error correction, and the array is designed not to trust them. Unlike SAS, error correction on enterprise SATA is (or should be) done by the controller, for example by chucking in an ECC sector every few sectors, independently of any striping. That's why you get 1/9 less usable space with enterprise SATA.
is a monitor stand that stands on the floor and works like a balanced arm lamp. That way I could get rid of my desk, and work either in a low chair xor standing up.
Second best would be one that clamps to a desk, but like the lamp, can stetch *below* its mount point.
My money is waiting.
The 2.5" drives are the gem in this range. No readily accessible useful info on which form factors are getting the upgrade to be seen on the WD site. Instead, we are informed that WD drives provide better traction.
Still, it's not as if magnetic disks are under any competitive pressure, so everything'll probably be okay.
I was away all last week. As soon as I arrived overseas, My Blackberry lost its personality and became a consumer device. I had no mobile email or data, only GSM.
Then the third party Internet provider at my hotel puked and stopped giving out IP addresses for three days. Now I really had no mobile email or data.
Roaming on my Android smartphone cost me €15/MB. I was reminded, poignantly, of precisely how little Google sees fit to cache locally, and how much update chatter it donates on my behalf. Thanks Google, my carrier loves you too.
Was I in Madagascar? Malaysia? Mali? No, friends, I was in Mountain View, CA.
I've looked at clouds from both sides now.
Sir, you have excellent taste.
Which would be *easier*? Building your i7 or buying the cross bike? Now replicate a million times.
The PC 'industry' bleats about falling sales, but until they learn to sell like the bike industry sells, they get negative sympathy from me.
I'm trying to build a machine right now. Impossible to find what I want. I'll probably stick with what I have, and buy that steel disc cross bike for winter.
Losers.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
