more and more people are asking me why gone have an additional distributed object caching system since even Microsoft started now a distributed / replicated caching system. I will try to answer this question in my personal point of view.
There are several reasons available, in comparison to MemCached we have a system which is written 100% in managed .net code and we started our first tries to run shared cache on Mono with a success (this version is currently not available - local development path - I will add more info about this in a different post) . Over the past few years MemCached approved itself over the biggest systems around the world and consider there is a great, active and helpful community behind it as long as people using it within Technologies like PHP, PHYTON and RUBI (and some more). Its quite clear (to me) they will use "only" MemCached and nothing else - nor Microsoft or Oracle will be able to create a solution for this communities.
At the time I started indeXus.Net Shared Cache there were nothing out there which was written 100% in .Net managed code, open source and FREE! There are several companies like Scaleout Software or Alachisoft they addressed a .Net solutions but the fact is as good they are - they cost money! In the meanwhile there are even starter kits available from various companies.
The first goal of Shared Cache was to give a solution for those who decide to use Microsoft .Net technology. The current status showed us that we able to compete in performance matters with MemCached.
Above comparison output I received from somebody else so I can't guarantee for it but on my environments results are quite similar.
What happens now since Microsoft announce it's Velocity project. I have contacted Microsoft Velocity Team with the question:
- Q:What kind of licensing will stay behind Velocity?
- A:The licencing terms have not yet been finalized. We hope to address it by CTP2.
Velocity's schedule based on the June's Tech-Ed 2008:
We hope Microsoft is going to implement the following requirements:
- No additional charge for Velocity
- 100 % managed Code
- They deploy at least pdb files
- Ability to run Velocity under MONO
- It will be at least that fast as Shared Cache and MemCached / Cacheman
I'm pretty sure most of above requirements they meet up until CTP 2 in October 2008.
Posts
2 comments:
Regarding whether Velocity will be free, is this really what you want to encourage? Please consider the effect it has on the marketplace for other vendors who have invested years in developing this technology. If you want angel investors and VC's to invest in Microsoft-related technologies that drive new innovations, then it's unwise to kill off the market for others just as significant customer traction emerges. While in the near term you benefit from a free platform feature, in the long term you suffer from the absence of a healthy, competitive marketplace that provides the best innovations and the widest variety of offerings.
Bill, you work for ScaleOut, i'd say you're biased, probably because your product is under featured and over priced. Not so over priced as nCache (those guys are out of their gord). The fact is, this is a feature that should have been part of the asp.net framework for a long time. If the niche providers that did enter the market (few because most people saw this coming) had decided to price their product approriately rather that try to ransom an ongoing percentate of scaling revenue then Microsoft probably would not have persued Velocity. License fees on par with the likes of SQL server itself are insultion considering the FRACTION of effort required to create the product.
Post a Comment