subscribe
« On Student Nature | Home | Identity and hypocrisy »
On platform lock-in
By Jay Savage | September 14, 2007
I’ve been mulling over David’s recent (ok, not that recent but RL’s been interfering) post on vendor lock-in and encouraging more PC use in his Mac-centric office. The premise is that having multiple OS will make your company more flexible and better able to respond to potential disasters. The example he gives is viruses. I don’t buy it.
First of all, the insurance policy isn’t worth the cost and hassle of supporting a mixed environment. Mixed platform file sharing and authentication is still hairy, and the people who can make a mixed network run smoothly demand, and rightly so, about 1.5 times the salary of monoculture techs. And at the end of the day, when disaster strikes you still only have somewhere between 10% and 20% of your machines up and running. For most businesses, that’s as good as none. Even 50% isn’t really enough capacity to effectively conduct business, so the money spent on the contingency plan is essentially money down the drain. Sure, a person’s machine may work, but if the people she collaborates with don’t have access to their files, if the people she relies on for reports and data mining don’t have connectivity or database access, or if the department fileserver is down, the fact that she has a bootable box is inconsequential. Buying all your employees large-capacity smartphones so they have backup access to email and important files is cheaper–and better insurance–than forcing some of them onto alternative platforms.
Second, the real concern isn’t employee workstations, it’s data. Unless the data center has redundant backups on servers running multiple OS, you’re still lock in for all practical purposes. If you lose the file server, you lose everything, no matter what your people have sitting in front of them. And the same is true of files on employee machines. It doesn’t matter if the person next to you has a PC if your files were on a compromised Mac, and vice versa. Files aren’t commodities. Simply having access to other files on other machines doesn’t solve the problem of a wide-scale outage.
For an internet services firm like David’s, of course, it’s important to have a bunch of machines running multiple OS to test usability and browser compatibility, but that’s a different question from disaster recovery. It can also be largely accomplished through virtualization.
The argument that Microsoft is still the de facto standard is more compelling, but I don’t really buy it. I didn’t understand it a decade ago, and I don’t understand it today. Data networks are about, well, data. As long as I can open your files and you can open mine, who cares whether the red “x” at the top of my windows is on the right or the left? If businesses want security, they should be looking at open file formats (and filesystems), not OS. Worry about vendor lock-in in for data and applications, and the platform will take care of itself.
Of course, that is one definite advantage of David’s plan: physically running multiple OS means his applications have to be platform-neutral, and probably rely on open–or at least widely-supported–file types. And that is a disaster plan I can get behind.
Topics: tech |
September 14th, 2007 at 2:46 pm
Hi Jay,
Thanks for the reply. Though I do want clarify my post somewhat. The issue of having computer diversity being “insurance” against Viruses was applied to my days as a currency options trader for the bank. And although I understand your position stated below:
“Mixed platform file sharing and authentication is still hairy, and the people who can make a mixed network run smoothly demand, and rightly so, about 1.5 times the salary of monoculture techs. And at the end of the day, when disaster strikes you still only have somewhere between 10% and 20% of your machines up and running. For most businesses, that’s as good as none. Even 50% isn’t really enough ”
This doesn’t hold up for a trader’s risk management. As financial markets move every second of almost every day. I needed constant reports on what our exposure was. In fact, had every PC melted down in our bank. That single Mac could have kept us a float and operational in London, Toronto, and Singapore for at least 24-48 hours. Meaning, I could continue trading (over the phone) and managing the Bank’s risk. We would eventually break down as the PCs were also responsible for payments and mark to market. However that initial 24-48 hours could translate to 10s of millions of dollars for my desk alone depending on the market conditions at the time.
My main concern now with my development company being majority Macintosh is not just viruses, but the unknown risks I can’t possibly understand due to my level of knowledge of a computer’s OS (does not knowing what I don’t know sound too Rumsfeld-ish of me?), as well as the closed thinking across various platforms which you mention.