- It exists too many vendors and solutions, I counted recently around 20 of them. This is a comment we must consider with the number of deals especially big deals as object storage is built for capacity and secondary storage. In other words, the market is not so large for object storage as customers can do things differently with good results.
- The Open Source accelerates the pressure and solutions from (by alpha order) OpenIO, OpenStack Swift, RedHat with Inktank Ceph, Sheepdog, SwiftStack, soon Minio, just to name a few, make life a bit complicated for proprietary software vendors.
- HTTP interfaces on top of NAS or application servers is very often largely enough.
- The Scalability justification for object storage is overkill as projects can be split in domains with gateway, proxies... that masks this segmentation. Again it works well.
- Object Storage is just a Data Storage Platform and even if people could have considered the beauty of that in the past, this is for a few quarters now attacked by new vendors who build and delivers still a Data Storage Platform but now rich set of embedded data services. That makes a serious difference.
- Cloud Operators swallow everything sooner or later, they have the keys: time and money except for things that must be kept on-premise within the enterprise perimeter so everything related to confidentiality requires a compliance dimension.
- Hybrid model seems to be the preferred model at least from vendors' perspective as they must limit the erosion of the full Cloud approach. We'll see.
- Recent acquisitions have a strong impact. It remains more object storage vendors than the number of potential buyers. NetApp picked Bycast in 2010, RedHat chose Inktank in 2014, HGST (WDC) selected Amplidata and a few days ago IBM announced Cleversafe acquisition. Of course I have to include Dell here with EMC with several object storage solutions in the EMC portfolio. So who is naked now ? HP, Oracle, Cisco, Seagate and for how many object storage preys ? Caringo, Cloudian, Compuverde, Scality, Tarmin... I'm really convinced that Cleversafe did a good move at the right time for all reasons mentioned above and also to avoid the dilemma to stay independent for greedy reason.
- Let's assume that big vendors will do their market and some object vendors stay outside. You get the point it will very hard for such players to be considered and chosen by users with all the pressure coming from big irons. These vendors will start to be out of the game rapidly.
- End to End solution from big vendors: IBM will have it, DDN, Hitachi, NetApp have it already, Dell has and will be stronger, HP has it but needs actively a good object storage platform, Oracle is similar but still Cisco is naked and Seagate has plenty of holes even more than WDC/HGST.
- Management in some object storage companies made mistakes for different reasons: lack of storage industry expertise, utopia in the product capabilities, too greedy and too pretentious and above all disconnected from the reality. It's good to have a vision but it has to be the right one.
Oct 19, 2015
The object storage segment is under a complete redefinition and the pressure increases every day on remaining independent vendors. Let me explain why.
Posted by Philippe NICOLAS à 2:36:00 AM