A thing I'd like to see: regional data centers

There is one thought that hit my mind when I read again the news about the possible opening of the server side software of Lindenlabs: it would really, really nice to have regional datacenters at least. Ok, having the possibility to host the servers on your own would be even better, but at least this step would be really nice in addition to the two datacenters in the USA.

Why? Simple: physics at work. What we all dislike is the almighty lag. Though the internet has become very fast, the connections between the continents are very fast, too, they still cannot beat physics.

But how does the lag involve physics? Well, even if the internet is fast, the maximum speed data can be transferred theoretically is light speed (roughly 300.000 km/s). You cannot cross that border. This means, the longer your connection gets, the longer it takes for the data to reach you. That's quite simple physics.

The time packets need in the network to reach their host is usually measured in milliseconds. When people play at LAN-parties they usually have ping times < 20 ms, so that's a very good time. Of course, when the way of your connection is, let's say, 10.000 kms long, the information alone needs about 0.033 seconds to reach it's target - one way, roughly estimated, that's still less than from Paris to San Francisco, for example and this still doesn't take into account the latency that several routers are adding up on the way. Of course, this means also no satellites!

But when your customers come from many different countries it's sooner or later better to have a regional data center in that region to give them better access to the data, which means less delay and a much smoother and faster experience. Since in the top 10 of the origin of the SL players 43.24% are coming from Europe (more than from the USA), the installation of such a regional data center would be the next logical step for LL to give customers better service.

I mean, when you look at other popular online games, they usually have one-two server locations in the USA (West- and East-coast), one in Europe and one in Asia. That's exactly because of the physics I explained earlier, and that's why I'd like to see the openings of those regional data centers. The company needs to move where its customers are.

Of course, when then again they are going to open up the server side part and you can host a simulator on your own, this would not be really necessary anymore. You just get a rental machine somewhere, where it's near your customers, get it up running and that's it. Well, we are going to see where the future leads us.

Downtime again and fun with Lindens

Another unplanned downtime, again. That's a good time to move on the voice beta grid and have some fun. Today there were many people at Ahern and Morris, so it's always good to move to another sim.

I was at Pulveria, there was a Linden and some other guys, but not so much as in the laggy welcome area and he spoke to us and with us. No better way to meet the Lindens in world than on the beta grid at the moment.

So... one asked the question, why a region sometimes gets better when restarted. Seems it depends on the neighboring regions, when you're restarting a region it could be moved dynamically to another CPU, so that's why it can become better, again. Interesting detail in the architecture.

Sinners Paradise - open again?

Today I tried to teleport in the sim of Wales. Yes, that's the sim which had Sinners Paradise in it and closed in January for the public due to massive and repeated griefer attacks. There's now also a new, adjacent sim in the making, Welsh Falls.

Well, I tried the teleport, it worked - interesting, since it was closed for so long. Seems it is open again for the public, the entrance area has changed, landscape looks more or less the same like before, now I wonder if this is wanted, and if it is really back in business again, how long it takes before this place kicks ass again.

Here's a little photo of me made today sitting in the entrance area of the teleport at Wales. Enjoy.

Eric Kintz from HP on Second Life

Eric Kintz, the vice president of global marketing strategy and excellence for HP, is running a very critical blog entry about second life. Taglein: Top 10 reasons as why I still need to be convinced about marketing on Second Life.

While he raises right points there, there's one point missing: this is going to be the next big thing[tm], if done right, the start of something new in that kind of matter, if done right, so getting experiences in SL will help you adapt later to other virtual worlds more easily.

Another roundup

Since I've got again too much ideas for blog entries in my mind and don't want to post them all as separate articles, I'm doing another roundup. Here it comes:

Well, that's it for the moment, more to come shortly.

The new search in town and the big privacy issues it implies

There is a new search tool available as beta for Second Life, being developed of the Electric Sheep Company at the moment and open for public now since a few days ago.

What's this thing and how does it work? Well, the searching part of it is a bot named Grid Sheperd, that runs from sim to sim, scans the objects of a sim and connects them with avatar names. The part available on the world wide web at the moment is a search interface, that enables everybody to lookup the items owned by the avatar. It is only showing the items of the avatar though that are tagged as "for sale". So the Sheeps are telling us: it's adding value to all of use, since it's automated and it's opt-out, if you don't want those infos available at all. They're very excited, announced the availability of this service here, described the options and doings of this search bot in this FAQ and an progress update in this blog entry. Important: opt-outs are now taking effect immideately!

Heck, I don't care about that! This thing is invading my privacy, plain and simple. I am going to ban the bot on every parcel I can and I am going to spread the word around and this bot makes me really, really angry, and I mean it!

So, what's wrong with this toy? Many things.

My few main reasons why I took an instant hate about it are:

  • Not everybody is going to know it is existant or how to ban it or can ban it.
  • On big flaw in the approach: it is opt-out! Sheesh! This thing should be opt-in only! I want to be asked before, not afterwards, it's that simple.
  • The technology used for this machine works and generates a big database. While per default only items tagged for sale are going to show up, I can visit the Sheep Island and make all of my items public available, if I wish. This means: it can be used to generate a big database with nearly all the objects I've placed in world somewhere. Bummer. Even if they don't use it to leave those items in the database, the technology is ready. Ready for what? Well, who guarantees me that not another person or company is going to use this to really scrape all data available on the main grid about me, generates files about me and is going to sell them to companies interested in my data for whatever reasons? This is a really, really big, bad breach of privacy! And this is also in my opinion a flaw in the architecture of SL and/or the handling of bots by Lindenlabs. This incident just pointed to a whole can of worms unsolved, we should be therefore thankful for it. Not everybody would announce such a thing in public, and that with good reason.
  • They're claiming it's adding value to the community. Well... perhaps to a small part of the community, but in whole this approach is very arguable. It would have been better if this bot was opt-in only from the start, but even then it's troublesome.
  • If I publish a website, I expect it to be indexed by a search engine. I can adjust the behaviour with a file named robots.txt, I can restrict access on my own and so on. When I open a store or build my house in SL, I don't expect such a thing happening at all. And if I don't read about it, I might never know it's there.

So, what's the conclusion of it? SL has lost some more of its innocence, if you want to protect your data, you must get your ass off and do it yourself. Since this bot is still in the beta, the default policy could change - perhaps. Then again, perhaps not. But if someone really wants to scrape the data of the grid about you, well, the technology is ready, and he's going to do it silently. You're never going to know what hit you until it is to late, then. And, believe me, there are many compromising objects around an avatar could place in his home which can backfire on your real life - badly. Very, very bad.

It would be even more easier for Lindenlabs, to do this on their own, since they're sitting on the whole database and just need to run some queries. Easy. But this would ruin their reputation at all, many people would perhaps threatening them with law suits, so this is not going to happen and we should trust them on that matter. Besides, you've got a contract with them.

And how is this new feature being reviewed by the community? That's really even more surprising: most are happy with it and welcome this! This seems for me that many really don't have a clue about what this actually could mean for them.

Some examples are:

Of course, such a tool has never only people supporting it; someone who got the whole shebang right and has the right views about it in my opionion is Prokofy Neva. He wrote actually a whole bunch of articles about it on his own blog and on the Second Life Herald and I must say, I share his views, that's very rare. Find his article about the search engine itself on his blog here, about the need to regulate the usage of bots (I wonder if it is possible) here and his article about it at the Second Life Herald here. He also wrote an open letter to the CEO of Lindenlabs, Philip Rosedale, in which he raises many valid points.

By the way, not only I've got a problem with this approach, according to another article of Prokofy Neva at the Second Life Herald Anshe Chung thinks this is a violation of her covenants and has banned the bot on her whole continent Dreamland. I can understand that move.

Well, what's left to say is: I don't think that ECS is evil. In my eyes they took the wrong approach to this delicate kind of matter, which is now backfiring great lengths on them. Backfiring how? Well, I know many potential investors, who are angry about it and for sure are now not going to give them orders for in world buildings and such, in contrary, even try to convince other people not to give ECS orders at all.

Another important thing is that this would have happened sooner or later, because Second Life allows this data to be scraped without much efforts at all, perhaps it already has happened many times and is still happening, done by others, and we just don't know it. Such things also have happened with different social network on the web many times, so better watch your steps. SL is like the wild west there, it seems - if you want security, your on your own - mostly. Better get used to it that an avatar is something in a database.

If you want to read more about that privacy in general, I suggest you visiting the Electronic Frontier Foundation (EFF).

Scaling Second Life

In the Avastar #14 there's a comment of Gwyneth Llewelyn about the scaling of Second Life on page 6. She's claming that one possible move to relieve the grid of stress would be to store the texture data, which seems to go into the hundred of Terabytes (wonder where she got this number from, I thought all data is around 34 terabytes at the moment according to this article, so I stuck with this number, so hundred of terabytes is wrong on a great scale), from outside the grid, perhaps even allowing users to store them on their own servers. While I don't see the textures stored on the own servers, because of possible protests, she continues.

She's stating that if that move occurs the 2000 servers of SL would be able to hold 20 million simultaneous users and this could be achieved within a month with the work of one developer!

Personally, I doubt that - really. I know they're going to switch from their own data protocol to HTTP somewhat this year. Serving a texture is a quite simple task - provided it's stored in a simple filesystem and not in a database. Storing textures - binary data - in a database system is always very dumb; the clever way to do it is to store the filesystem name in the database only, since the database is much, much slower to the task and adds far more complicity to it.

And, of course, the textures and so on are really not stored on one hard disk, but I guess on a logical volume or clustered filesystem. There are enough technics around to do it under Linux.

I personally think that serving the textures is not what strains the servers, because they're loaded into the cache and that's it, then. It's more computing the viewing range, making database querys about the prims (if stored in database) with their textures, it is interpreting scripts, computing the locations of all the avatars, doing the physics stuff and so on. And that's why I think even if the textures are located somewhere else that the technic now available on the main grid is not up to the task to handle 20 million users at the same time