With yesterdays outage of Gmail for two hours, the discussion about when to use SaaS or host application can erupted again. Should you run others SaaS or is it better to run the applications in your own datacenter. This is also interested for Google Wave, since most users probably will use Googles hosted version.

I will not start guessing if the internal IT organization is better at running their hosted Wave server or if Google is better at running their complex Wave servers. At least Google has the resources to get the system back up running fast. It would require a large IT organization to get systems back online in just 2 hours.

The outage gave information on how depended we are on Gmail or mails. When (if) we convert to Google Wave this dependency will get even bigger, because more communication will go that way.

As a person I can live without email, Twitter or Wave for minimum 6 hours. For communications part it can probably be down longer, but it is more the data I have stored that I need. I have hotel reservations, which would be quite frustrated not to have when I need to find the hotel that, I booked. It will be inconvenient not to have access to some data when I need it, but it can probably be managed.

For organizations downtime is more costly because employees cannot work, or have to perform other work than processing emails. There is of cause difference in how dependant each user is of mails, some just checks mails few times a day while others work with the mail constantly. Some of the work can be perform using snail mail or just save the mails in text file, which could be create to mails later. All the processing towards the ERP systems can be performed as they always have, so there is no need to retrain people.

If Google Wave is down, when organizations start depending on the service, the work is stopped again. Wave currently does not have an offline version, and it will be strange to have offline waves. So the works will have to reply or communicate using email instead.

With mail people has to enter data into their ERP system manually. With Wave it is possible to let the wave interact with the wave directly. This means that employees cannot process the orders offline, after they start to use Wave.

If Google Wave’s is down, and then organizations how host their own Wave server, they can still work. The problem would be that there will not be any communication outside the organization, this might not be possible because of Googles users how cannot interact with the waves anymore.

The whole problem is the more dependants we are on a technology the more vulnerable we get for failures. I don’t believe that many organizations can do a lot of work when their ERP system is down, we might see the same for Wave systems.

Related posts:

1. Google Wave as a replacement of Notes databases I was talking a person how was using Lotus...
2. Workflow applications in Google Wave One interesting aspect of Google Wave is the way...
3. Using Groups in Google Wave Using wave in an organization can be a little...

Related posts brought to you by Yet Another Related Posts Plugin.

Anil Dash wrote that the Wave as a too complicated protocol to implement. Anil writes that nobody but Google will create a waveplatform because it is too complex to implement.

“But the fundamental Wave protocols are, I fear, a bit too complex to ever be fully and correctly implemented by anyone other than Google. Interoperability is likely to be a challenge that plagues the platform for its entire existence. In short: It’s likely that nobody will ever build a fully-compatible clone of Wave that competes with Google’s own implementation.”

Yes it is more difficult to write a federate wave application, that to use the Twitter API for communication. But don’t we (as humans) have the ability to implement complex protocols.

How about SMTP created in 1982 with an RFC, a document on 67 pages. I don’t think that I’m able to get an SMTP server up and running on a weekend, as is Anil’s time for a easy implementation. You also needed to considere how TCP/IP worked before you could implement the SMTP protocol. The wave protocol draft specification is 24 pages (when I copy the text to Word), so it is not longer. If the complexity together with XMPP is larger I cannot say.

When talking about open specifications and complexity. Then the winner must be OOXML or ODF, I don’t know the size of this specification but people are fighting to create some useful specifications that others will implement. I don’t know how many will implement the standards Microsoft and OpenOffice, but it is open if people want to implement features into them.

Agreed we will not see a lot of providers of Wave servers developed by third parties. But some have already started with developing clients as open source projects and there will probably be more projects. We will probably also see some software providers, how will implement and sell their own version of wave servers/clients, so you can have the privacy of on premise solutions.

Another area where there are specifications and multiply providers is in the J2EE world. Sun created a speciation of how ie. J2EE 5 should work and then multiply developers created their version of an application server. In this area we have Sun, JBoss, SAP, (Oracle) and countless other servers.

Sure I would like the Wave protocol to be as easy to program as Twitter, but then we have not created a solution which could revolutionize communication.

Related posts:

1. Wave News for 7 september 2009 For this week there has been some interesting Wave...
2. Etiquette on Wave to is too complicated One of my favorite blogs on Google  Wave (The...

Related posts brought to you by Yet Another Related Posts Plugin.