Wednesday, July 3, 2013

#62: The Mythical Cloud: Don't Believe the Hype

A while back, as of the time of writing, Game Trailers released a news story about a game named Darkspore, published by Electronic Arts and developed by Maxis, makers of The Sims and SimCity. The game was newsworthy because after many months of problems with it, rendering it nigh unplayable, Steam has officially delisted it, making it unavailable for purchase. This is interesting because the issues have nothing to do with the game itself. In fact, the real issue is that the EA servers are not operational. That is correct: In order to play Darkspore, even in single player, customers had to connect to EA servers. You see, EA had cited that the servers were needed for the betterment of the experience, and not as a form of DRM. It is this particular issue that I wish to talk about: the cloud.
Lately, especially with the dawn of the next generation of gaming consoles, cloud gaming has become a serious talking point for both Sony and Microsoft. Quite a few next-gen games, most notably Titanfall from Respawn Entertainment and Ubisoft's upcoming games The Crew and The Division, also made it perfectly clear that they will require the use of cloud-processing in order to function. I have heard a lot of people in the industry praise the advent of cloud-processing. After thinking about it, I am not convinced that such an innovation is healthy for the industry. However, I cannot just say that I think it is bad. My task is to argue the point to those who do not share my view, which I intend to do.

The biggest point someone prosecuting cloud-processing in a video game can make is that utilizing it is another way to mandate that players of the game are constantly connected to the internet at all times. In other words, regardless of any potential benefits, it is another form of always-online DRM. In order to actually use cloud-processing services, it is necessary to connect to the servers where the calculations being remotely handled for the purpose of streaming inputs and receiving outputs. Since the odds of any given person playing the game in the same room those servers are located in are <1%, this can only be reliably done via the internet. Going further, this connection must be maintained in order to continue to make use of the cloud for offloading calculations, because otherwise there is no way to transmit data. The unfortunate reality of this necessity means that cloud-processing will always demand that users get online and stay online, giving publishers and developers a very convenient excuse to implement always-online policies. We have seen this is the past with 1 very infamous case study.
Of course, I am talking about SimCity (2013). Also released by Maxis, SimCity (2013) was actually the fifth installment of the series with the same name. Unlike previous installments of the city-building simulator, this game featured an ability to communicate with other players' cities and share resources between them. Sadly, this otherwise interesting feature came with a cost. EA decided that to facilitate the sharing of resources, all saves had to be uploaded to their cloud servers, with no copies on a given user's computer. In order for this to work, EA mandated the players be constantly connected to these servers. When people cried foul at this, Maxis claimed that due to the way the game was programmed, it was literally impossible to add this new feature without also including always-online. This was later proven false by a simple modder who allowed the game to function perfectly fine without any sort of online connection. Though EA and Maxis deny that the system implemented was for DRM to this day, most of the people who saw it unfold would be hard-pressed to accept that for truth. The notion of utilizing the cloud was quite obviously another way of sneaking DRM into a game that does not need it.

Another issue with cloud-processing in video games is the sheer impracticality of its use. Though I spoke of this last week when talking about the Xbox One, the issue with maintaining servers applies equally to cloud-processing. Needless to say, without servers to offload calculations onto, it is impossible to actually do any offloading. Creating and maintaining said servers cost money, which is already a dwindling and precious resource in the industry. Since many people in the industry lament a lack of profits due to a variety of reasons, it seems foolish to knowingly forge an unwritten and unspoken contract with consumers to keep and maintain servers so that a given product remains playable. Though not living up to their side of the contract is certainly still legal in this case, it has a way of tarnishing a publisher's reputation, and lowering consumer confidence in future products.
Another key practicality issue is that there are only so many calculations that can be offloaded to the cloud. Ignoring the issue wherein many people do not have reliable internet connections, there is only so much data that can be transferred through wires and even the air itself. Bandwidth is very much a finite resource, so it is necessary to limit the amount of data that needs to be transferred between cloud servers and the machine playing a given game. As a result of this limit, things like high-definition graphical data are pretty much completely out of the question. While I do not consider my own internet to be particularly terrible 70-80% of the time, there is absolutely no way I would ever be able to stream HD graphics through my connection. I can barely play YouTube videos at 480p. Other types of calculations exist on a spectrum of practicality, so at best cloud processing can really only be used as a supplement and/or for games that are not very system intensive. With this in mind, claims from Sony regarding Gaikai, which is their reported “solution” to backwards compatibility, and Microsoft having “infinite power with the cloud” seem dubious at best. To best utilize cloud processing would require an intelligent, and nuanced, approach that minimizes the amount of data streamed through the internet. That notion contradicts claims made by both major console manufactures.

The last, and maybe most significant, issue that cloud-processing presents is that over the long term it will result in a lack of longevity for video games that use it. Like I said in my previous point, utilizing this new computation technique requires servers that must be maintained. Although it will probably take a very long time for most games, eventually the time will come when a business decision gets made. It will be decided that the costs to keep servers up and running outweigh any benefits of keeping them, so they will shut down. Like in the aforementioned case of Darkspore which started this article, cloud-based games that suffer this fate will be permanently shut down unless publishers are generous enough to put the tools out there for consumers to make servers for the game. This results in making games that, unlike every other entertainment product out there, have a finite, if unknown, shelf life.
To be fair, this is a very hard case to make to publishers. After all, whether or not a game works 10 years from now does not really have a noticeable impact on profit margins. Also, making games like this enables them to charge years later for higher definition remakes and ports of those same games, giving them a financial incentive to make games that will expire in some way, shape, or form. However, in the future, it will be necessary to have these products available as a way to study and learn from them in many the same way people learn from old books or movies. If the servers no longer exist for these games to be played, then they will forever be lost to history. This is a problem that has yet to truly be solved, and even services like Steam will need to eventually face it. Although the thought sounds laughably absurd, there will eventually become a time for each of these companies to forever close their doors or be merged into another. This new technology is ripe for abuse in this regard, and that is something that can be frightening to many.

On that note, I want to make very clear that this article was not intended to be an alarmist piece on cloud-processing in video games. All I wish to do is inform you, the reader, that there are still many concerns that must be addressed before it becomes a more viable model. There are positives uses for it, too. After all, Steam, Playstation Plus, and Xbox Live users have already become accustomed to the benefits of storing backup save files on the cloud. For games like the aforementioned Titanfall and The Division, which have made it clear to consumers that they are exclusively online multiplayer games, cloud-processing is a perfectly viable tool for offloading some extraneous calculations away from the console/computer.

It only becomes concerning when we see cloud computing in areas where it does not necessarily belong, like in the case of Darkspore, as the problems then outweigh the benefits. There is also the opportunity for a nuanced approach, using cloud-processing only for the multiplayer components of a game, not affecting the single-player portions. It can, in fact, be a benefit to gaming. However, it needs to be done smartly, else many other problems are born as a result. It is a difficult balancing act to make, and I am honestly not sure that major console manufacturers, game developers, and publishers are able to do it. Who knows? Perhaps I will be wrong. Nonetheless, make sure that you are informed about the technology before you pass judgment one way or the other.

No comments: