On Friday February 26th, Lab Gab, the live streamed chat show hosted by Strawberry Linden on all things Second Life, returned with a show of two halves.
Featuring guests Grumpity Linden, the Lab’s Vice President of Product and Oz Linden, the Lab’s Vice President of Engineering, the first part of the show took a look at the latest of the work to migrate Second Life and all its services to running on Amazon Web Services (AWS) hardware and infrastructure and attempted to address questions forwarded to the Lab by Second life users.
|Table of Contents|
The show was also an opportunity to say “farewell” to one of the leading lights at the Lab – Oz himself, who is retiring from the company and from full-time work as a whole – after more than 11 years with the company.
The official video of the segment is available via You Tube, and is embedded at the end of this article. The following is a summary of the key topics discussed and responses to questions asked.
Please be aware that as some topics were touched on more than once during the conversation, the notes blow collect references together, rather than presenting them in chronological order. However, where relevant, time-stamps are provided.
On the Status of the AWS Migration and the Future
- All of the services related to Second Life were transitioned to running on AWS hardware and infrastructure by the end of December 2020.
- The last aspect of the core work was the removal of all of the Lab’s own hardware and equipment from the Arizona co-location facility that had been hosting Second Life, which included the shredding of 10,588 hard and solid state drives to ensure data security.
- The majority of the work went a lot more smoothly than had been thought might be the case, however, there are some services that have given rise to some problems that are still being resolved.
- Chief among the latter is the Land Store, which was once again turned on ready for use on Thursday, February 24th.
- Map title generation has also been a issue sine the migration, but work is progressing on fixing this.
- [9:09-11:34] A core issue with the Map tile generation failure lay in the fact that the code had not been touched in a “long, long time” – so long, in fact, that the code isn’t geared to rendering mesh objects, hence why they can look so abstracted on a map tile.
- In terms of the current problems, the code made a lot of assumptions about the architectural environment in what it was running, assumptions that are no longer true with the move to AWS.
- The current work is focused purely on getting the service to generate Map tiles one more, without making any additional changes to the code to account for things like rendering mesh objects correctly or addressing other bugs.
- Most of this work is now thought to be complete and Map tiles are now being generated as they should. however, there is some work to be completed on stitching tile images together when a user zooms out on the Map.
- There is a project to improve the overall appearance of Map tiles, but this was put aside in the focus of migrating to AWS, but will hopefully be picked up again at some point in the future.
What is Next?
- While the physical migration of Second Life services from a proprietary environment to AWS is complete, the Uplift Project work is not, and so will continue to be a focus of engineering efforts.
- In particular, the immediate focus is on optimisation work, which encompasses:
- Optimising the performance of the various series on the new hardware / infrastructure.
- Optimising (for the Lab) the cost involved in running within an AWS environment.
- Fine tuning systems and operations within the new environment.
- Working to leverage the new hardware options and infrastructure presented by AWS to favour Second Life as a product running in that environment.
- In this it should be remembered at the initial migration work of getting Second Life transitioned to AWS was devoted purely to taking all of the SL services – front-end simulators, back-end services, middleware, web properties and services, supporting tools, etc., – from the proprietary environment in which they had always run and just getting them running on AWS in what was called a “lift and shift” operation, whilst making as few changes to any of the services as possible.
- With the “lift and shift” aspect of the work completed, the engineering team has turned its attention to gathering data on exactly how the various services are running in the new environment and understanding where opportunities for making the improvements noted above may lie, and how they might best achieve them.
- In this, the Lab now has much improved service monitoring tools at their disposal, and these are now allowing the initial work on tuning performance on key services to be made.
- Two practical benefits of the move are:
- Regions running on AWS can run more scripts / script events in the same amount of time than can be achieved on non-AWS regions.
- The way in which simulators are now managed mean that LL can more directly obtain logs for a specific region, filter logs by criteria to find information, etc., and the entire process is far less manually intensive.
Will the Migration Mean lower Prices for Users?
- Sort answer for the foreseeable future: no.
- There has been an idea circulating among users that running SL on AWS is “dramatically cheaper” for Linden Lab; but this is actually not accurate.
- Prior to the migration, all of SL and its services had been running on LL’s own hardware for which there had been no capital expenditure for years, and which had completely depreciated.
- The move to AWS represents something of a new capex spend, increasing the Lab’s costs [although it is not unfair to say that the capex involved is liable to be significantly less over time than repeatedly buying-in new server clusters to allow SL to run on more modern systems].
- Rather than presenting LL with reduced costs, the move to AWS is designed to:
- Present the company with far broader options for delivering a more performant and capable service to users – although as noted above, it will take time for all of this to be delivered.
- Improve the overall longevity of the Second Life service through the noted performance improvements and access to better hardware and infrastructure services.
Second Life Mobile App Status
- Mobile has taken longer than expected to bring forth, for two primary reasons:
- The first is that while the initial release will be more of a communications tool, considerable foundational work has been put into ensuring the app can be encompass a lot more functionality than that in the future.
- The second has been that as a result of testing by Apple, the Lab has been forced to make changes to the way in which chat works.
- These changes will, in time, be filtering through into the viewer as well.
- They should actually make chat more reliable in the future.
- No commitment as to when the app may be more widely available.
Other Technical Questions
- [11:38-17:47] There have been numerous niggling issues of late: further issues with search (e.g. avatars failing to show in search), profile issues, etc). When are these likely to be addressed? Should users report bugs then find?
- Whilst the majority of the migration process did go smoothly, there have been glitches, and the Lab is working to address them alongside of working on the performance, etc., work mentioned above.
- There are a lot of aspects of SL built on old technology, so there is an expectation that, over time, and as things can be looked at, not only will niggles go away, but software and capabilities as a whole can be made a lot more stable and resilient.
- Bugs should most definitely be reported using the SL Jira. Information on how to file bug reports (and feature requests) can be found here:
- [17:55-19:18] Will capabilities that were being worked on some time ago (e.g. 360 snapshot viewer) ever be completed?
- The migration work has demonstrated what can be achieved with a tightly defined set of goals and teams focused solely on those goals.
- This is an approach Grumpity would like to carry forward, with a commitment to review current and past projects to determine what might be required to bring them to completion (input, time, resources, etc), and then make decisions from there.
About the Lab’s New Owners
- Linden Lab put itself up for acquisition in 2020 (see Linden Lab announces it is to be acquired, and Linden Lab’s acquisition: sundry thoughts & speculation), a process that was completed at the end of the year (see: Meet Linden Lab’s new board of directors).
- Since then, board member Brad Oberwager has joined the Lab’s management team in his capacity as Executive Chairman).
- The overall feedback that the new board and their staff are enthusiastic about the platform, and have brought a lot of energy into the company.
- Mr. Oberwager has apparently made an impact with the management team by not only putting forward ideas but responding to ideas by saying, “Do the right thing for the residents” – a phrase Ebbe Altberg brought into the company when he took over as CEO.
Looking Back at Oz’s Time at the Lab
The latter half of the programme looks back over Oz’s time at the Lab and provides him with the opportunity to discuss what attracted him to Linden Lab, the nature of his work, why he regards his time with LL as potentially the best job he’s ever had, and to discuss his post-retirement plans and answers various questions.
Rather than offer a summary of this part of the show, I encourage people to listen to it in full, as it really is informative and enlightening, particularly if you’re not familiar with Oz’s work, his teams, or the Lab as a whole.
The end of the show sees Strawberry teleport Oz to s special in-world retirement party, where the teams reporting into him and other LL staff have gathered to wish him well. This again a touching and moving tribute that says so much about Oz and the high regard in which he is rightfully held, and should be seen without input from the likes of myself.
For my part – and because Oz has been both a direct and indirect influence in my SL time – I’d like to just repeat what I wrote a few days ago on reading of his upcoming retirement:
For my part, I cannot claim to know Oz as well as I would like to – but I’ve always found find his enthusiasm for Second Life never to be anything less than totally honest and infectious, and his high regard for users utterly genuine and sincere.
As such … I’d like to take this opportunity to offer him a personal and public “thank you” for all the times he’s provided me with insight and / or encouraged me to get involved in various projects, all of it has been greatly appreciated. I am, and will be, genuinely saddened to see him leave the Lab; we are all losing something in his departure, and the void left will not be easy for the management team to fill.