20 February 2007 - 16:51Don`t go against free-ware
I was reading this entry on Cameron Purdy`s blog about Tangosol`s Coherence being among the 10 best products for your next gen data center when I remembered Hibernate`s clustered cache being released. When I read about it I wondered how would Tangosol compete with this free product. The answer, in hind-sight, would be that Tangosol didn`t compete with Hibernate (you cannot compete with a free product) but exited that market and entered the data management market. Tangosol, I presume, continued to fine-tune its cache engine, but focusing only on fine-tuning for differentiating your product from a free product is not a long-term strategy: it simply places your product into the very high-end part of the market, sometimes against your will. Just compare Ferrari’s market to Toyota’s market.
Tangosol`s main differentiator is spelled out in one word: management. Hibernate will not manage your 200+ data cluster, will not allow you to implement transparently various provisioning strategies, etc… Coherence will give you what Hibernate`s clustered cache has to offer plus a some other cache-related issues and a ton of management options. These options are what you pay for when using Coherence and less its caching engine (which I assume to be among the best in the world). I am sure that Coherence`s caching engine is a lot better than Hibernate`s but I think its real value lies in cache management. If you have a different opinion I would like to hear from you.
P.S. Come to think about it, you could add some real value to Hibernate’s clustered cache by slapping a good management console on it. And charging for it, of course.
Later edit: Following Cameron Purdy`s comments it looks like I was wrong about assuming that Coherence serves the middle of the market for replication/caching. Coherence seems placed in the high-end part of that market. That being said I think my reasoning is still valid, so I’ll keep this post. Feel free to comment.
3 Comments | Tags: Development, Favorites, IT in S-E Europe, Management
21 Feb 2007 - 14:57
Thanks for your kind comments, they are appreciated.
21 Feb 2007 - 16:50
When we designed and released Coherence, we purposefully picked extremely difficult problems to solve, because we did not want to compete in a commodity market, and we definitely did not want to compete with open source software. It is difficult (near impossible) to “compete with free”, and (speaking personally) I will tend to prefer open source solutions myself, so how could I expect others to act differently?
I do hope that business models around open source will mature and succeed, but I also hope that there will be enough technical creativity to attack and solve “that next generation of hard problems,” which will likely require an investment-heavy approach that traditional software business models can effectively support.
In our case, we may solve a problem that fifty banks would each pay one million dollars to solve, but (if it were open source) they might only pay ten thousand dollars to support it (i.e. “professional open source”). In a large scale grid of thousands of servers running risk calculations, most banks already build 80-90% of the software themselves, so they are hardly scared by the idea of having to figure out and support yet another open source project. On the other hand, when we show that we can drop their 2-month calculation to under one hour, they are more than willing to pay to use the technology.
While it will be a few years before this becomes clear to the industry as a whole, I can already see it today in our own market, which (at least at the low end) is rich with both open source and commercial solutions. It is evident that, even with great amounts of funding and very low expectations on returns, the open source businesses will likely not be able to survive. Very simply, the economics of support-based businesses requires mass market adoption. Even JBoss (a very successful open source business in many ways) was losing money, and they were as close to “mass market” as any professional open source company in the Java space.
Our result has been seven straight years of doubling+ of our revenues, which is a good growth curve, and one capable of fueling our R&D (which is quite expensive). Yes, we have been very fortunate, but it was not entirely accidental: We focused on a high-value, technically “challenging” (i.e. near-impossible) area of work, and as a result have been able to build a good name for ourselves, a defensible market position and a great business.
This approach is something that I hope to see replicated (no pun intended): Pick a super hard problem that lots of people have but no one has bothered to solve, and get to it! The future is waiting.
Peace,
Cameron Purdy
Tangosol Coherence: The Java Data Grid
21 Feb 2007 - 18:55
Cameron, it looks like I was wrong about considering that Coherence serves the middle of the cache market, from your comments it is definitely focused on the high-end market.
What I tried to convey in my post was that if the hard problems get solved and their solutions become commodity (very big IF I have to admit) one way to escape this market is by focusing on how to manage these solutions.
I would assume that thru its relationships with its customers Tangosol acquired insight into data-layer management and not only into replication technologies and I think that this insight is very valuable and is a key asset on its own.
Thanks for taking the time to comment and all the best,
Cristian.