Ongoing Updates from the Grid from FJ Linden

FJ Linden here, to report on the latest Ongoing Updates from the Grid.

As I promised in my first post, this will be a regular monthly communication to keep all of you up to date on our efforts to improve grid stability and reliability. I’m finishing up my 3rd month at the Lab and have some significant progress to report.

I’m happy to report that we have an approved plan to move away from VPN reliance. We’ve finalized a design and chosen facility and equipment partners to build and deploy a private fiber optic ring to interconnect our datacenters. “LLnet” will be the designation of our private network and we have established an aggressive timeframe to activate it. I’m pushing hard to bring LLnet online by the end of this year (’08), and begin a phased migration off of the VPN’s immediately after. Given the amount of traffic to move, I would estimate completion of this project by February or March of โ€˜09 at the latest. So we have a light at the end of the tunnel on one of our biggest stability issues.

In addition to freeing us from our dependency on VPN’s, LLnet will bring additional redundancy, as we will have a means to keep our datacenters from being isolated, if a Level 3 outage were to occur. We’ll also be building a fully routed architecture that will allow us to deploy BGP and finally add redundant Internet providers and improved traffic performance. This is a major infrastructure investment, but one that is long overdue.

I’ve included some timeframes in this posting, but do so only to show that we are pushing as aggressively as possible to get the network operational, if timeframes slip I’ll make sure to communicate that as well. I think itโ€™s important that we start making hard commitments and push aggressively to meet them. I know everyone is tired of talking about the VPN problems, and you can count me in that group! We’ve now got a solid plan in place and are implementing!!

I also wanted to highlight a few additional infrastructure projects that are targeted for completion by the end of 2008.

  • HTTP Dataserver: The goal of this project is to migrate all C++ mysql traffic from mysql protocol to http(s). Not only will this further free us from VPN dependency, it will also allow us move off of MySQL wire protocol over the WAN and better enable tracking and monitoring of queries. We’re about 50% of the way through this effort and expect to be complete by the end of the year.
  • Agent Inventory Services: The goal of this project is to provide a RESTful web services interface to the agent inventory. This is one of the projects in flight to address the numerous issues that Residents experience with inventory, and is designed to simplify communications and provide better reliability. Phase 1 of this project will be finished this quarter and should have a positive impact on perceived inventory loss due to unreliable messaging.
  • So, I’ll continue the monthly updates and, while I’m sure there have been individual issues since my last post, we are making real progress towards our primary goal of grid stability and reliability. As always, the proof is in our execution of these projects, not the promises.

    This entry was posted in Announcements & News, Bugs & Fixes, Development, Security, Service Status. Bookmark the permalink.

    78 Responses to Ongoing Updates from the Grid from FJ Linden

    1. Ciaran Laval says:

      I can think of thousands of $50 reasons why the timing of this post won’t be appreciated.

    2. wandasfashion says:

      This sounds very promising. If the abolishment of VPN will allow the grid to stay calm and working under higher concurrency loads than the 65k it takes now at best before crippling, it will mean we can finally deal with the neccessary growth in userbase.

      Given the last 18 months expansion in land area as opposed to actual users, that would be a most welcome thing. Especially for those of us that rely on userbase growth for their in-world businesses.

    3. Nice post… Bad timing……..

    4. bbot says:

      Ah, some tough talk from LL at last. Good to hear.

    5. Cincia Singh says:

      Is there ever a good time to post?

    6. Vivito Volare says:

      Before you get tons of backlash from previous announcements in this thread, let me say: Sounds great, good news.

    7. Sin Trenton says:

      Say, do you chaps communicate with each other at the Laboratory? Ok, we know about your communication skills in this direction, but.. internally?

      The Agent Inventory Services sounded interesting, in another situation.
      Best of luck, anyway!

    8. Elfod Nemeth says:

      Good news, hopefully this will allow better sim crossings, though doesn’t seem related.

    9. Jessica Hultcrantz says:

      One word: Finally ๐Ÿ™‚

      Reading about the redundancy and abandoning of the VPN’s makes great news in times of a big dark depression all over the world.

      Hopefully this will improve the experience for all abroad as well due to a smoother packet traffic and better connectivity. Private fibers are never wrong for big data flows. And redundancy helping against outages as well, about time as you said ๐Ÿ™‚

      And thanks for telling it with some technical details as well. Appreciated.

      Good luck with this important (and overdue) work!

      /me gives TJ a cookie

    10. Coventina Dalgleish says:

      Very good news, if it works. I will wait to see the proof of the pudding though as many past promises were just words. Not a very good time to announce as all it will beget are complaints from the poorly thought out increase in the open island pricing and the reduction in gross revenue it will create.

    11. Great news! And I’m glad to hear some specific details about how you guys are addressing the stability and scalability issues, and this clear communication is much appreciated. ๐Ÿ™‚

      (And yeah, as others have said, you’re likely to get a lot of flak just because of the post that preceded yours. D’oh!)

    12. Nyoko Salome says:

      gahhh!! sorry fj for all the others who can’t direct their comments to the many others already in forum… ;0 and yes, we all know about strategizing p.r., ok enough of that!! ;0

      ON-topic;0 – this all sounds super-excellent and moving forward to a better stabler grid experience… in a necessarily ‘lossy’ environment, we still cannot go about having inventory being subject to that ‘lossiness’ ๐Ÿ˜‰ very important to everyone, no matter what they are doing! keep it up and looking forward to what the grid becomes…

    13. Magical Beaumont says:

      Anyway, these are great improvments to the grid which should be honorred.
      The additional charging of allready existing contracts for openspace sims, which would be broken with this new charging rule, will be solved by lawyers around the world and not by us here. So, lets concentrate on the promises for grid improvments ๐Ÿ˜€

    14. Danball Tureaud says:

      Ok, this is really great. Fiber = less lag and much lower ping times between servers. I wonder what the next stability increase will be? Maybe they plan on using Sun servers (if they don’t already)

      On the previous blog post, people seem to have lost the ability to do mathmatics. $75 to $125 is not a $60 increase, thats $50, but the 66.6667% percentage is correct

    15. Christos Atlantis says:

      This is a very good move and it does show that SL is moving in the right direction. I can’t wait to see the improvements ๐Ÿ™‚

    16. Samsam oh says:

      YAY!!! ๐Ÿ™‚ My head of fluffy flexy prim hair won’t lag the sim so much!!! Now I can get that sexy flexy chest hair to! Thank you great linden god!

    17. iliveisl says:

      well thank you for the soon to come improvements. i love sl and appreciate all that you and everyone does on my behalf, but as stated above, the timing is poor

      i’m sure it will all work out (no other options)

    18. JJValero Writer says:

      FJ Linden, i have one suggestion. Perhaps may be possible to reduce traffic to the asset Dataserver and eliminate many problems.

      It consists of adding to each simulator one MySQL that take charge of storing all those objects and textures that are in the SIM.

      When one avatar create one object, the object only is store in the asset database if avatar take one copy or take the object to inventary.

      The main idea is cached in a more aggressive, but in a way that there are no inconsistencies with the asset server. It might be worthwhile to do the experiment in one SIM of beta sandbox.

      This text has been written using a translator. Sorry if it seems written by a Martian.

    19. JZ says:

      I agree with most… Keep up the good work. I for one, look forward to a more stable Second Life. Enough said, now back to my cave.

    20. August Hayek says:

      Mr. FJ, thank you very much for your post. It sounds like good news. I imagine that will be a boon for throughput and stability.

      When you say “private fiber optic ring” do you mean dark fiber, or are you getting some MPLS goodness from someone?

      What does BGP buy you? I always figured you had assets in non-portable netblocks at co-los all over the world. Are you going to light up an AS for purposes of having better control over the routing of inbound access?

      Thanks again. It warms my heart to see this sort of information to the customer base.

    21. G.W. says:

      That’s great news but I have to agree with the others this is terrible timing. A question though FJ, is the eventual goal of the new inventory system essentially to allow full item control within your inventory from a web interface and if so are there plans for a similar system for transactions? I was going to ask about friends as well but I figure that would be beating a dead horse since it’s asked so much with no result.

    22. SeanMcPherson Senior says:

      Interetsing. My other post was removed. I’ll rephrase.

      “Do you anticipate this project becoming suddenly easier with the projected changes in LL server loading (as large numbers of servers currently hosting openspace sims are now able to be powered off and/or have their network changed underneath of them), and do you think the month-to-month capital LL will shortly free up (by not having to buy new servers for several months as they re-deploy all the servers about to be abandoned due to the openspaces pricing change) will enable you to buy whatever gear and fiber is needed for this project?”

      How’s that work?

    23. AWM Mars says:

      Any stats on what this disolving the VPN reliance will mean in the longterm to users?
      Getting away from bottle knecks is not only welcomed, but also about time. From what I see, most of the hinderance on potential growth, would be the limitations of interconnectivity between the various asset nodes and database timings. My question is specifically, what will the overall experience for users be?

      Is this a move towards the desireable autonomousy for users with regard to their inventory and IP rights? Making them not just transportable between grids/platforms, but have localised backups? I sincerely hope so.

      Its a ray of sunshine atm, between the gloom clouds that surround the previous post of price increases. However, prices are rising to reduce the loadings on the system, isn’t this new direction going to make that almost a mute point?

    24. Danball Tureaud says:

      #20 Gigs Taggart, $125 is not 150% of $75, it is 166.67% of $75
      Try 75×1.6667 = 125

      If it were 150%, 75×1.5=112.5

    25. Amilie Anatine says:

      very happy about improving performance, something to look forward to. thank you! ๐Ÿ™‚

    26. Raudf Fox says:

      I would like to add my voice to the, “This sounds great! I’m very glad to see this post,” group. I’m very glad you’re keeping us posted on your progress.

    27. Isadora Fiddlesticks says:

      Hi FJ,

      Thanks very much for the much needed update! I for one welcome it especially after the decision to increase OS prices. Great news despite the development.

    28. Isadora Fiddlesticks says:

      er, Great news despite the EARLIER development regarding OS…sort of balances it out, huh? end the day with good news.

    29. Andromeda Quonset says:

      Interesting. I must be a little dense here, but could somebody please define for me what an “http dataserver” is. So far, I haven’t found it defined ๐Ÿ™‚

    30. Katt Linden says:

      Well, this is FAR from my area of knowledge, but this Wiki page does explain the various elements of the “system” that can cause disruptions, including dataservers:

      http://wiki.secondlife.com/wiki/Service_Disruptions#Dataservers

    31. Andromeda Quonset says:

      Thank you, Katt. That is a nice overview. Databases are not my area of expertise, but is one of many things I’ve studied recently. But among my references, most everything points to some level of an SQL server, with the differences just being in how other software talks to the server. I will continue looking ๐Ÿ™‚

    32. Aphrodite Tagore says:

      FJ, it sounds like you are someone with a strong network reliability background and dedication to world class performance. I can almost remote view your Ishikawa diagrams, and busts of Demming and Juran in your cube. Top notch talk. Hopefully followed by top notch performance.

      Sad that one of the first friends and coworkers I have had inworld, PonygirlSarah Clapper must reduce her provision of interactive, social roleplaying content due to the inadequatrte means LL has chosen to address the OS overload issue. This content, with its moderated and organized recreations, has grown to benefit many RPers in world, at no small expense to Miss Clapper.

      I would hope to point out that, as an end stage of improvement, hopefully holographic storage of assets will be achieved, as, in addition to perceived loss due to inadequacies in asset cluster pipes which produce “percieved loss”, there is true, actual unrecoverable large-scale asset loss, and has been for as long as I have played. Holographic, or at least, fully redundant real time non-volatile storage of the huge asset data base must be achieved. Holographic architecture is no doubt superior in grave outage situations, and careful cost analysis would prove more financially expedient in the long run. This is where the CERN grid architecture also falls apart, in that data is still handled linearly and discretely, rather than as a holographic image recoeverable from any sufficient numbers of grid nodes, although CERN is relying on simple redundancy instead. That is now a suspect mode, as holgraphic memory is now achievable.

      I wish you all the best in your efforts, as you are the first Linden to express an understanding of major issues and a concise plan to eliminate the difficulties we experience. I expect that within a rasonable time frame, and with your assistance, we Residents may experience 6-sigma quality here. In satisfaction of customer desires only, does a business generate huge profit, the natural result of efficient process and superior customer satisfaction.
      Can you imagine the wonder when people, skeptical or previously disaffected with the basic playability issues are told through word of mouth “Hey, go create a free account and look in SL…it is now perfect.”

    33. Eva Ryan says:

      WHEW…good thing I didn’t buy into that Open Space broo-ha-ha…

      …oh, wait, that was the other blog post…

      Well now, good new…bad timing. Sorry can’t offset the difference. But, it’s good to know that the folks at Linden Lab are hard at work stablizing the grid. Glad to see that VPN is being gone away with. The backbone structure is showing it’s osteoporosis, and the doctor said that surgery is needed…and how. But, remember, folks…surgery always must come with a recovery period. Be prepared for that.

      I now return you to your regularily scheduled program.

      -Eva Ryan (That’s Eva Ryanโ„ข )

    34. Alicia Sautereau says:

      btw, isn`t this what happend at AOL and you came from there to be CEO here?

      hmm funny coincidence

    35. Well, for all you NAYSAYERS out there, you better hold on.

      FJ, I really mean this, but keep up the work. I know all of this will take a while to accomplish, but seriously, you can sleep when you are dead. There are24 hours in the day, let’s make use of them all!

      I really can’t believe I’m saying “Good Job” to a Linden. Must be my meds or perhaps a blog entry that reads like a status report, which I like!

      OK, FJ, you promise to keep us updated yes? I look forward to hearing more about how things are progressing in this area; keep up the good work!

    36. Simon Kline says:

      Wow great news, expecially about the fibre optic networking! The grid monkeys and LL are amazing considering the number of them and the number of us they’re supporting! I found this article on how they tracked down an issue to a faulty memory ic in a router once, well worth a read to appreciate what they’re working with:

      http://wiki.secondlife.com/wiki/User:Poppy_Linden/Network_Errors_and_Data_Loss_2008-01

    37. well I think the grid would run much better without all the disturbing residents

    38. If I understand well, instead of communicating between datacenters through the internet (VPN technique), LL will have its own physical network.
      I hope this will solve the rash of gratuitous crashes which are to occur recently (often several an hour), where the viewer seems to function normally, but is actually disconnected from the grid.

    39. Fox says:

      Sounds interesting, thx for sharing with us,
      this with mono and all other things in the works, looking forward to it, keep up da great work !

    40. Sorry for being off-topic in my previous comment (61). I thought I couldn’t comment on the other post anymore.

      It really is a good thing and we all do hope to see stability and performance improvements because of this in the future. The timing of this post could have been a LOT better though.

    41. This looks awesome, although the timing is really upsetting. I’d say that this news would have positively impressed the SL community much more if you delivered it in a different moment.

      I’m looking forward to a better blog and communication management, as Robin announced a few days ago.

    42. Maggie Darwin says:

      Thanks, FJ. These moves all sound well-grounded technically.

      Of course, “by the end of the year” always means extra holiday work for the worker bees. And change always carries with it the risk of instability during the transitional period.

      When you begin to get the creeping feeling that one or another element of the plan is starting to look too aggressive, bear in mind that the holidays are a peak shopping time and a bad time to be unstable.

      Thanks for keeping us informed of your plans.

    43. Saii says:

      The good & the bad it is to me.
      Nice something is done about redundancy, but it’s sad to see mySQL is still used.
      mySQL was and still is for small to medium sized business, since the underlying technology is not “matrixed” based. In other words, not suitable for a system where everything runs on database operations.

      Yeah mySQL is popular, because it’s free, easy to setup and can scale up indeed. But did you ever do a complex performance comparison?

      Improvements… good; You can polish a Lada until it shines brightly in the sun, but it will never be a Ferrari.

    44. Thanks, FJ. This should be a great help.

      Has anyone looked at why the viewers are ignoring the cache on startup? This is a relatively recent problem and everyone seems to be experiencing the same thing, even after a graceful restart. When I come back inworld to my last location, everything is grey. I have tried setting my cache at max and at 3/4 max with my draw distance set very low (64m), yet everything seems to be refetched. Often, this occurs at a very low bandwidth of sometimes as low as 15kbps (I am on an 800kbps cable connection). Redrawing the entire area takes as much as 15-20 minutes.

      Using the local cache would vastly reduce the load on the backend. I rarely leave my world, so there is really little need to go back to the asset server.

      I realize this is not your area, but we should find out why the load is so high to begin with. The concierge has tried moving our servers from Arizona to Texas, but it made no difference.

      Thanks for improving the infrastructure. Unfortunately like others, I believe we are approaching the end of our SL. But I still have hope that it will not die out completely. Perhaps what you learn in improving SL can help those worlds that do survive.

    45. Gil Druart says:

      Thanks FJ – please ignore all the off-topic twittering – those architectural changes sound immensely encouraging and just as encouraging is the fact that you seem to have this project firmly by the scruff of the neck. Keep it coming.

      [Is tempted to cut’n’paste an article about BGP exploits in order to foment panic in all the airheads, but decides that while it might be funny it wouldn’t be helpful]

    46. Cheta Torok says:

      While I can’t comment on the effectiveness of the upgrades, their promise is appealing. To me it appears you are making serious efforts to address system stability by improving the messaging and cross data center communications. While I know some of the changes do cause some latency issues, I would bet they are smaller than the issues with the existing systems. And more robust for that matter.

      As for the timing, I love the timing, ASAP please ๐Ÿ™‚ I for one will welcome a more stable environment. I deal with people having inventory issues from time to time and have lived through the data center communication disruptions, this year having some doozies.

      Thank you for your continued efforts to make the platform for SL, the place I know and love, more stable and as a result, more enjoyable.

    47. Rylan Oldrich says:

      It is refreshing to see someone at Linden that seems to understand that stability of the basic grid is the single most important issue that has been ignored. I hope you are more successful that your fellow lindens at accomplishing something that really benefits the users (make that paying customers). It is a shame that your efforts will be too late for so many of us who tried to make SL something special, who have been betrayed by the openspace pricing. While I applaud your efforts, for Linden Labs as a company its just too little too late.

    48. Pingback: Ummm, Is THAT Copyrighted, Too? « Common|Sensible

    49. Evad Babii says:

      Excellent news FJ – I’m sure all the corporates and ‘partners’
      will be really pleased, seeing as they will eventually be the
      only people using SL the way things are going.

      Residents? – those that are staying might enjoy it for a while –
      that’s if there’s any left by the time you have it implemented.

      As some have said – too little too late for many ๐Ÿ˜ฆ

    50. Lucien Franciosa says:

      Wow, I guess once people pick up tomatoes to throw they just can not put them down! Sorry FJ.

      My comment on what you said is… say what? I admit to not knowing a VPN from a ATP from a Supercalifragilisticexpialidocious . I appreciate you gave us a lot of details and a timeline and I saw the word ‘stability’ in there, but effective communication should always consider the audience and my guess is most SL users are not Networking experts. Please give up front next time something the rest of us can understand rather than blah blah blah stability blah blah blah. Just a suggestion. My guess is less tomatoes would be thrown about the other post if there was something here the majority could sink their teeth into. As it is, it appears to be a ‘let’s cover up that other thing with something, anything, positive’

    51. Katt Linden says:

      Please keep your comments here on topic.

      Thanks.

    52. Sedary Raymaker says:

      Cool, this will help the mainland run better once LL pushes private islands out of the market.

    53. Alexandra Rucker says:

      @ 51

      Katt – that was an extremely unpopular decision, out of the blue, with no apparent consideration for people who already bought Openspaces, especially hours before it was posted.

      Just how earth did you people EXPECT they would react, anyway?

    54. shawn suisei says:

      ok everyone, lets just retract our claws for a moment about the “previous posting” before this one and look at this in the light of the whole big picture.

      Stability and a well functioning grid are key to retaining residents and bringing in new residents (and these are key to any inworld business). As a long time SL’r (well into my third year here) I have seen massive improvements in the way SL runs and I give a ton of credit to all the hard working people at LL who have accomplished the small miracle that is Second Life.

      And to all of those who are whining about the lindens bring in to much land and loseing the value of their land, let me just say that i support anything that increases the size of the Second Life world and
      really have little patience for those who want to restrict the worlds growth so they can line their own pockets. (do i hear oil cartel? – the screaming of restrict the flow of an asset so everyone has to pay more and a few landlords (oil barons) get wealthy). Anyone who views virtual land in a world were more land is always being created as an “investment” really doesn’t understand how a commodity is valued.

      And to the Lindens, as a long time supporter of your efforts let me tell you in no uncertain terms that your radical, across the board price increase on OS is stunningly stupid.

    55. shockwave yareach says:

      I rather thought that my comment was on topic. How many years of “we’re gonna fix it” have we heard? How much confidence in LL do we have today?

    56. Walentine Andersson says:

      Katt, maybe you and the rest of Linden Labs should stop hiding then and ANSWER the topic that people seem to care about right now!!!!!!!

    57. Raban Laborde says:

      it’s a realy bad timing for this news..

      but, indeed, it’s a VERY good news!

    58. shockwave yareach says:

      A SIMPLE SOLUTION

      First, everyone who bought the voids at 75$ a month, get a choice: keep it 75$ a month with a 12 person limit on the parcel, or they may sell it or abandon it. Grandfather people for the next 8 months with a notice that the rate will rise to 100$ mid next year.

      Second, for the next 3 months, allow people to abandon Voidsims without losing everything else they own.

      Third, make all new void purchase 100$ across the board and make it clear that 12 people are all that can be in it at a time.

      This will prevent 60 avatar dance clubs from crushing the voids and dampen the fury your actions have created. A hard limit on the number of people in a void should prevent clubs and their script loads from remaining in voids. And keeping the prices as is for at least 8 months, people don’t feel like you’ve ambushed them. For those who still feel ambushed and want to dump their voidsims, allow them to do so without any penalty – throwing so many people and their thick wallets out of SL is not a wise business move.

      Much distrust have you created. It is up to you to attempt to make it right again. Sure, it’s your system and you basically can do what you like. But it’s OUR money and time, and we are likewise free to take it and go if you behave like greedy mortgage lenders.

    59. miffed says:

      After January 1st, I won’t care. I’m leaving my OpenSim and not spending one more cent in SL.

    60. Sorry to be off topic, but the previous blog passed in a very short time. Delete this if you want, but at least please know the customer feedback.

      Most open lands I know are properly used (few prims, few people in). I however witnessed those open sims becoming suddenly very laggy, because somewhere else a club or whatsit was abusing of the resources. So the problem DOES happen commonly to legitimate users. But the solutions are clear:

      1) hard limit the number of people, prims, scripts… in open lands. You know that there are abusers, and abusers abuse, even in a democracy.
      2) allow people with four open lands to be on the same server.

      The solution you retained (suddenly making open lands much more costy) is e very displeasant one, it is making everybody pay for the abuse of a few, and for the lack of appropriate enforcement. Please reconsider this bad move, which cancels many other good moves.

    61. Zemaitis Flatley says:

      Texture loading too slow,and a lot of Teleport error
      Plz fix this version….

    62. Ann Otoole says:

      The fiber rings are good. Can’t get to a million concurrency without it.

      I strongly recommend you get a real global scale middleware solution in there for transaction management.

      The question now is if a million concurrency is even logical to consider at all anymore. I put my thoughts about it over on my blog.

    63. Buckaroo Mu says:

      Oh lord, I have waited for this news for YEARS. Ever since they split the datacenter between SF & TX, I’ve been encouraging them to buy up some dark fibre & get rid of this VPN stuff. And BGP! No more relying on the world’s least reliable ISP, Level3! Oh Glory Day, It’s come at last! (well, it’s coming.)

    64. Stimpy Tripp says:

      Hi:
      The switch from VPN hardware to dedicated fibre sounds exciting to me. I’m glad you’re doing it on an aggressive timeline, and it looks like some thought has been put into both speed and redundancy.
      As we all know, when the network is slow or flaky, SL becomes slow and flaky so this is an important step in combination with the better reliability that I, for one, am seeing in the client.

      Thanks! Stimpy

    65. Danielravennest Noe says:

      Dear Frank,

      Since you are infrastructure kind of guy, can you put some numbers on just how much more resources 4xOpenspace sims uses than 1xregular sim? From our customer viewpoint, it looks like the number of server CPUs being used (one) is the same, and the prim allowance (15,000) is the same. The total grid has 30,000 sims, so that means the average number of avatars per sim is under 2.5 even at the highest concurrency ever reached. Therefore it is simply impossible for ALL the openspaces to be loaded up with people, there just are not enough of us. The average openspace must be showing something on the order of a couple of people in it.

      I think most players can understand paying for what we use. To me though, the appearance is that openspaces had 1/4 of the server and prim allocation, and were priced at 1/4 of the full island rate (75 vs 295/month).

      If the reality is that an empty sim (just bare land and water, no user prims) uses a significant portion of total resources (server cpu cycles, networking, etc), so that running 4 sim processes per CPU core eats a lot of the capacity, then allowing 3750 prims/openspace was a mistake. But tell us in actual numbers whats going on, and where the load comes from. And here is the important point:

      *Charge us according to resource use*

      If its prims, scripts, avatars present, or some combination, then set the pricing according to what is the actual cost in resource use. And give light users the option to have lower limits and pay less accordingly.

      Its simply not fair to say after the fact “we didn’t think you would use all the prims we gave you” after doubling the allowance and dropping the price. That shows an amazing lack of understanding of human nature, and lack of load testing a maxed out openspace server to see what would happen.

    66. amilie anatine says:

      hey FJ, tell everyone to clean out their trash and lost and found folders, that might help performance.

    67. My Two Cents says:

      Don’t come down on Kat about the Open Space stuff. The whole thing has been blown waaaaaaaaay out of proportion. Including clogging the heck out of the JIRA because for some reason someone thought to use that instead of a stupid online voting poll or to use a petition site. The whole Open Space thing is stupid. I’ve seen tons of Open Space sims abusing the heck out of their prim limits using temp rezzers too many scripts and other stuff to get around the limitations.

      But opening up JIRA on it? Get real people!! It’s not going to help. This choice is based on economy. It costs money to run electric to the computers running these sims. So, as costs go up and the stock market crashes. Prices do have to rise. Everything costs money. Then you have the JIRA getting clogged up with whining and flaming and just insanity. Really if what… 50 dollars is such a big deal why buy it in the first place. It’s like moving in to an apartment sometimes when you renew the lease you have your cost of living go up.

      That’s what’s happened here. Cost of living has gone up because Open Space users have abused them. CPU’s can only stand so much and if they were running lower grade or even running more then a few on one server, then yes! Lag, complaints, and man hours on the Lindens side trying to track down problems. Then they find the problems! Too many Open Space users pushing the CPU above and beyond the call.

      If you don’t like the management then move! If you don’t like the policy then leave. The agreement says that prices can change without notice. You were given notice. Be grateful! Other companies who have done things like this up their prices with no notice. You just log in and notice that you have a higher price then last month.

      And about this post. I have seen VAST improvements in the stability of SL! And I’m very pleased about it. There are less outages and problems are getting handled ASAP! This post is all great news to me. And to all that think MYSQL is crap… Have you ever programmed? Mysql is great even on a business server running intensive programs. I don’t use anything else in all my projects because it has yet to let me down.

      Thank you and I apologize in advance for this rant. You can delete it if you’d like. I just wanted to add my two cents. There has been way too much cry babying about the price hike. Everyone just needs to get over it. Supply and demand and expenses. Electric is expensive!

    68. Phillip Vought says:

      FJ from a fellow network architect, congrats on seeing the easy win and on getting leadership to see it too..

      I have to ask if there is going to be a concerted movement to an SOA within Linden Labs.. An ESB with an SSL foundation and XML heart would certainly releave an enormous amount of strain on the individual database components and massively increase reliability..

      Seeing as you have a good working relationship with IBM already, are you looking at or even thinking about a move to SOA conceptual foundations such as MQ?

    69. Pingback: OpenSimulator « secondlife is psychic

    70. Ashelyn Dryke says:

      Another brilliant move by Linden Labs! I know I’m not technically a Second Lifeยฎ customer and can’t officially support the Linden’s efforts, but I for one am very excited at the expansion of the Second Life Gridโ„ข, especially for business applications! I’m sure this expansion will be very profitable and will certainly benefit Linden Labs. I hope every resident will support Linden Labs financially so it can continue to become more and more stable and become a truly awesome virtual business platform! Great job Linden Labs!

    71. ChatNoir Moonsoo says:

      First time I’m getting censored here..and that for asking a pertinent question (relating to contradictory statements from LL).

      Go figure.

    72. Alf Lednev says:

      Votes to delete My Two Cents Rant Business 101 Two cents, never tell a paying customer to leave, learn it. Katt why aren’t you or Missing Jack on the Forum answering the questions on the OS mess? The silence is deafening. all the promise of better communication, openness etc are shown to be meaningless at the first big test. Your losing ground, do something about. Once credibility and trust go, they are very hard to win back. And the OS forum farce is making a lot of goodwill is evaporating.

      As to the Grid, now, no one but the geeks will care until they know the pricing overlay Lindens will impose. Everything costs and Lindens only solution to users is charge more.

    73. Mills Gazov says:

      @70

      It is difficult to be able to support Linden labs and the Second Life Grid when Linden Labs make massive changes without warning that disrupt the SL economy as much as the recent price hike has.

      The change to your own fibre network has been long over due and will greatly improve the experiences of residents and i am looking foward to when it is completed, I just hope that you don’t mess up the switch over.

    74. Ashelyn Dryke says:

      @73

      I was actually being sarcastic. That and all the negativity aside, I DO think this is obviously a good thing. Better stability and speed is always good for the grid and for the residents. I’m just concerned that this isn’t specifically for the residents, but rather to make the grid more stable as a business platform. If more businesses hop on the SLGrid bandwagon, how much of this fiberoptic network do you think is going to be devoted to SL and the residents?

    75. Pingback: Openspaces: A Proposed Solution - SLUniverse Forums

    76. Pingback: The lie about land « Pandora’s box

    77. Pingback: Openspace announcement is up - Page 4 - SLUniverse Forums

    78. Pingback: M Linden: Second Life Update and Welcome to Howard Linden (aka Howard Look) « Official Second Life Blog

    Comments are closed.